system

Abstract

We provide the system. [Solution] A system for evaluating the reliability of news articles, comprising means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments.

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background Art]

[0002] Patent Document 1 discloses a persona chatbot control method executed by at least one processor, the method comprising: receiving a user utterance; adding the user utterance to a prompt including an instruction associated with a description of a character of the chatbot; encoding the prompt; and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior Art Literature] [Patent Literature]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2022-180282 [Summary of the Invention] [Problem to be Solved by the Invention]

[0004] At present, in order to evaluate the reliability of a news article, individual readers need to personally perform collation with primary information sources, fact checking, and comparison with experts' comments, and this work requires specialized knowledge and also takes time. In addition, there is also a lack of means for article writers themselves to objectively evaluate the reliability of their own articles. [Means for Solving the Problem]

[0005] This invention provides a system for automatically evaluating the reliability of news articles. Specifically, it includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. This allows readers to easily verify the reliability of articles, and enables writers to objectively evaluate the reliability of their own articles. [Brief explanation of the drawing]

[0006] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Embodiment 1 of Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1 of Form Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2 of Embodiment 2. [Figure 14] This is a sequence diagram showing the processing flow of the data processing system in Application Example 2 of Form Example 2. [Figure 15] This is a sequence diagram showing the processing flow of the data processing system in Embodiment 3 of Example 3. [Figure 16] This is a sequence diagram showing the processing flow of the data processing system in Application Example 3 of Form Example 3. [Figure 17] This is a sequence diagram showing the processing flow of the data processing system in Example 1 of the Form 1 when an emotion engine is combined. [Figure 18] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1 of Form Example 1 when an emotion engine is combined. [Figure 19] This is a sequence diagram showing the processing flow of the data processing system in Example 2 of the Form 2 when an emotion engine is combined. [Figure 20] This is a sequence diagram showing the processing flow of the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] This is a sequence diagram showing the processing flow of the data processing system in Example 3 of the Form 3 when an emotion engine is combined. [Figure 22] This is a sequence diagram showing the processing flow of the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. [Figure 23] This is a sequence diagram showing the processing flow of a data processing system in another embodiment. [Modes for carrying out the invention]

[0007] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, the labeled processor (hereinafter simply referred to as "processor") may be a single arithmetic device or a combination of a plurality of arithmetic devices. Further, the processor may be one type of arithmetic device or a combination of a plurality of types of arithmetic devices. Examples of arithmetic devices include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), and TPU (TENSOR PROCESSING UNIT (Registered Trademark)).

[0010] In the following embodiments, the labeled RAM (Random Access Memory) is a memory that temporarily stores information and is used as a work memory by the processor.

[0011] In the following embodiments, the labeled storage is one or more non-volatile storage devices that store various programs, various parameters, and the like. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disk (e.g., hard disk), magnetic tape, and the like.

[0012] In the following embodiments, the labeled communication I / F (Interface) is an interface including a communication processor, an antenna, and the like. The communication I / F manages communication between a plurality of computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (Registered Trademark), Bluetooth (Registered Trademark), and the like.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0014] [First Embodiment]

[0015] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0016] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0018] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0019] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0022] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0023] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0024] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0026] Next, the identification process performed by the identification processing unit 290 of the data processing device 12 will be described.

[0027] "Example of form 1"

[0028] One embodiment of the present invention is a system for evaluating the reliability of news articles. This system has a means for verifying the consistency between the text of a news article and its primary source using information from online sources. Specifically, it compares the facts and data described in the article with publicly available papers and databases on the internet and calculates the degree of agreement.

[0029] "Example of form 2"

[0030] Furthermore, the system of the present invention also includes means for fact-checking. This verifies whether the facts and information described in the article can be confirmed by other reliable sources. For example, it verifies whether the incident reported in the article has been reported in official police statements or by other media outlets.

[0031] "Example of form 3"

[0032] Furthermore, the system of the present invention includes means for cross-referencing with expert comments. This means that the specialized content and views described in the article are evaluated in comparison with the opinions and comments of experts in that field. For example, it checks whether the medical claims reported in the article are consistent with comments and papers from experts in the medical community.

[0033] The following describes the processing flow for each example of the form.

[0034] "Example of form 1"

[0035] Step 1: The system receives the news article text as input.

[0036] Step 2: Extract the facts and data contained in the article.

[0037] Step 3: Compare the extracted facts and data with publicly available papers and databases on the internet and calculate the degree of agreement.

[0038] "Example of form 2"

[0039] Step 1: The system receives the news article text as input.

[0040] Step 2: Extract the facts and information contained in the article.

[0041] Step 3: Verify whether the extracted facts and information can be confirmed by other reliable sources.

[0042] "Example of form 3"

[0043] Step 1: The system receives the news article text as input.

[0044] Step 2: Extract the specialized content and viewpoints mentioned in the article.

[0045] Step 3: Evaluate the extracted specialized content and views by comparing them with the opinions and comments of experts in that field.

[0046] (Example 1)

[0047] Next, we will describe Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0048] There is a need to quickly and accurately assess the reliability of news articles, but traditional methods are time-consuming to cross-reference and evaluate information, making it difficult to obtain reliable results.

[0049] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0050] In this invention, the server includes means for verifying the match between the news article text and primary sources using information from literature on an information network, means for analyzing the article content using natural language processing technology, and means for searching publicly available sources on an information network and comparing facts and data within the article. This makes it possible to quickly and accurately evaluate the reliability of news articles.

[0051] A "news article" is a piece of writing intended to convey information, providing details about a specific event or occurrence.

[0052] "Reliability" is an indicator that shows that information is accurate and error-free, and it is a standard for evaluating the legitimacy and credibility of information.

[0053] An "information network" is a communication infrastructure for sending and receiving digital data, including the internet.

[0054] A "primary source" is the initial source of specific information or data, and is a source that provides direct evidence or data.

[0055] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and it involves analyzing text and extracting meaning.

[0056] "Public information sources" are databases and documents that provide information that is generally accessible to the public, and are collections of information that anyone can use.

[0057] A "text similarity calculation algorithm" is a computational method for numerically evaluating the similarity between different texts, and it measures the degree of textual agreement.

[0058] "Evaluation results" refer to the outcome of an evaluation conducted based on specific criteria, and indicate conclusions regarding reliability and accuracy.

[0059] To implement this invention, it is necessary to build a system for evaluating the reliability of news articles. The user sends the text of a news article to the server. The server analyzes the received news article using natural language processing technology. This analysis utilizes natural language processing libraries such as "spaCy" and "NLTK". The server extracts facts and data from the article and searches publicly available information sources on the information network. This search utilizes academic paper search APIs and database APIs.

[0060] The server compares the content of news articles with primary sources obtained through searches. For this comparison, it uses text similarity calculation algorithms such as "Cosine Similarity" and "Jaccard Index." The server uses these algorithms to calculate the degree of similarity and generate evaluation results. These evaluation results are provided to the user via a web interface or API.

[0061] As a concrete example, consider a case where a user wants to evaluate a news article about "the global economic growth rate in 2023." The server extracts the statement "the global economic growth rate in 2023 is 3.5%" from the article. Next, it uses an academic paper search API to search for related papers using the keyword "2023 global economic growth rate." The server compares the content of the papers obtained from the search results with the description in the article and calculates the degree of similarity using Cosine Similarity. Finally, the server evaluates the reliability as higher if the degree of similarity is high and returns the result to the user.

[0062] An example of a prompt to input into a generative AI model might be, "To evaluate the reliability of the news article, please compare the facts in the article with online sources."

[0063] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0064] Step 1:

[0065] The user sends the body of a news article to the server. The text data of the news article is provided as input. The server receives this text data and prepares for the next analysis step.

[0066] Step 2:

[0067] The server analyzes the text of received news articles using natural language processing techniques. Specifically, it uses the natural language processing library "spaCy" to extract facts and data from the articles. The input is the text data of the news article, and the output is a list of the extracted facts and data.

[0068] Step 3:

[0069] The server searches publicly available information sources on the information network based on the extracted facts and data. This search uses an academic paper search API. The input is a list of extracted facts and data, and the output is a list of relevant primary sources. Specifically, the server performs a search using the keyword "2023 global economic growth rate".

[0070] Step 4:

[0071] The server compares the content of news articles with primary sources obtained through searches. The "Cosine Similarity" algorithm, a text similarity calculation algorithm, is used for this comparison. The input consists of the text data of the news articles and a list of primary sources; the output is a numerical score indicating the degree of similarity. The server uses this to evaluate the reliability of the articles.

[0072] Step 5:

[0073] The server generates evaluation results based on the calculated degree of similarity. These evaluation results are output as numerical values ​​and comments indicating the reliability of the news articles. The server provides these evaluation results to users via a web interface or API.

[0074] (Application Example 1)

[0075] Next, we will describe Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server," and the smart device 14 will be referred to as a "terminal."

[0076] In modern society, a vast amount of news articles are distributed via the internet, but these sometimes contain unreliable or misleading information. This creates a risk that readers may believe false information. Therefore, there is a need to quickly and accurately evaluate the reliability of news articles and provide this information to readers.

[0077] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0078] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, and means for comparing with expert comments. This makes it possible to quickly and accurately evaluate the reliability of news articles and provide readers with reliable information.

[0079] A "news article" is a piece of writing or news report created to convey information, and is primarily distributed through the internet or print media.

[0080] "Reliability" refers to the degree to which information and data are judged to be accurate and free from errors.

[0081] An "information network" is a general term for communication systems, including the internet, that are used to exchange data and information with one another.

[0082] A "primary source" refers to a document or literature that provides the origin of information or the first recorded data.

[0083] "Means of verifying consistency" refers to methods and techniques for comparing the content of a news article with information from primary sources to determine whether they are consistent.

[0084] "Means of fact-checking" refers to methods and techniques for verifying whether the information contained in a news article is accurate.

[0085] "Expert comments" refer to opinions and explanations provided by experts who possess knowledge and experience in a particular field.

[0086] "Methods for cross-referencing" refers to methods and techniques for comparing the content of news articles with expert comments to verify their consistency.

[0087] An "information terminal" refers to electronic devices such as smartphones and tablets that are used to display and operate information.

[0088] A "generative AI model" refers to an algorithm or system that uses artificial intelligence technology to analyze data and is trained to perform a specific task.

[0089] A "confidence score" is a numerical indicator that quantifies the reliability of a news article and is used to evaluate the accuracy of the information.

[0090] The system for implementing this invention uses an information terminal and a server to evaluate the reliability of news articles. The information terminal is an electronic device such as a smartphone or tablet, which is used by the user when viewing news articles. The server performs the central processing for evaluating the reliability of news articles.

[0091] The server receives the text of news articles and uses bibliographic information on the information network to verify their match with primary sources. This involves using APIs to access publicly available databases on the internet (e.g., Google® Scholar, PubMed). Furthermore, the server uses a generative AI model to analyze the content of the news articles and calculate a confidence score. This generative AI model is built using machine learning frameworks such as TENSORFLOW®.

[0092] Once a confidence score is calculated, the server sends the result to the information terminal and displays it to the user. If the confidence score is low, the information terminal displays a warning to the user to draw their attention. This allows the user to immediately judge the reliability of the news article.

[0093] For example, if a user views a news article about the effectiveness of COVID-19 vaccines on their smartphone, the server evaluates the article's reliability and sends a reliability score of "85 / 100" to the user's device. If the reliability score is low (e.g., 40 / 100), the device displays a warning message.

[0094] Examples of prompts for a generative AI model include the following:

[0095] "Analyze the content of the news article and compare it against the following databases: Google Scholar, PubMed. Calculate the article's confidence score and output the result."

[0096] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0097] Step 1:

[0098] A user views a news article on an information terminal. The information terminal retrieves the text of the news article the user is viewing and sends that data to a server. The input is the text of the news article, and the output is the transmission of data to the server.

[0099] Step 2:

[0100] The server analyzes the text of the received news articles. Specifically, it uses natural language processing (NLP) techniques to tokenize the article's content and extract important keywords and phrases. The input is the text of the news article, and the output is the extracted keywords and phrases.

[0101] Step 3:

[0102] The server uses the extracted keywords and phrases to refer to publicly available databases on the information network (e.g., Google Scholar, PubMed) and verify their match with primary sources. It uses an API to query the databases and retrieve relevant bibliographic information. The input is the extracted keywords and phrases, and the output is the relevant bibliographic information.

[0103] Step 4:

[0104] The server uses a generative AI model to calculate the confidence score of news articles. The generative AI model compares the content of the news articles with the acquired bibliographic information and evaluates the degree of agreement. The input is the content of the news articles and the bibliographic information, and the output is the confidence score.

[0105] Step 5:

[0106] The server sends the calculated confidence score to the information terminal. The information terminal displays the received confidence score to the user. If the confidence score is low, the information terminal displays a warning message to alert the user. The input is the confidence score, and the output is the display to the user.

[0107] (Example 2)

[0108] Next, we will describe Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0109] In today's information society, the internet is flooded with a vast amount of information, some of which is unreliable. Therefore, both information recipients and providers are required to quickly and accurately assess the reliability of information. However, traditional methods for verifying information reliability are time-consuming and labor-intensive.

[0110] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0111] In this invention, the server includes means for verifying the consistency of the information text with a primary source using information from a communication network, means for performing fact-checking, means for analyzing the information using an information processing model and generating relevant terms, means for searching external sources to verify the information, and means for analyzing the search results and determining the reliability of the information. This makes it possible to quickly and accurately evaluate the reliability of the information.

[0112] A "system for evaluating the reliability of information" is a device or method for determining the accuracy and reliability of information by using information from materials on a communication network against the text of the information and confirming its consistency with the primary source.

[0113] "Information on communication networks" refers to information sources such as documents, articles, and databases that exist on the internet and other digital networks, and involves using these sources to verify and confirm information.

[0114] A "primary source" refers to a reliable source that provides the origin of information or original data, and serves as a standard for verifying the accuracy of information.

[0115] An "information processing model" refers to an algorithm or program used to analyze information and generate related terms, making the content of the information easier to understand.

[0116] "External information sources" refer to information providers or databases that exist outside the system and are used to verify and confirm information.

[0117] "Means of analyzing search results" refers to methods and devices for evaluating information obtained from external sources and determining the reliability of that information.

[0118] This invention is a system for evaluating the reliability of information, with a server at its core. The server receives information input from the user and performs a series of processes to evaluate the reliability of that information.

[0119] First, the server analyzes the input information using a natural language processing library. Specifically, it uses natural language processing software such as "spaCy" to tokenize the information and extract important elements such as nouns and verbs. This analysis allows the server to identify the subject of the information and related words.

[0120] Next, the server uses a generative AI model to generate relevant keywords from the extracted information. Based on these generated keywords, the server searches external information sources such as the "Google News API" and the "Bing Search API" to collect data to verify the reliability of the information.

[0121] The server analyzes search results obtained from external sources and determines the reliability of the information. Specifically, it compares search results and evaluates whether the information can be verified by other reliable sources. This evaluation result is provided to the user to help them quickly and accurately determine the reliability of the information.

[0122] For example, if a user enters "an article about the announcement of a new technology," the server analyzes the article and extracts keywords such as "technology" and "announcement." Next, it uses these keywords to search external sources and retrieve relevant news articles and official announcements. Finally, the server analyzes the search results and reports to the user whether the content of the article could be verified by other reliable sources.

[0123] An example of a prompt message is, "Please check if the information in this article can be verified by other reliable sources." This prompt allows users to efficiently evaluate the reliability of information.

[0124] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0125] Step 1:

[0126] The user inputs information they want to evaluate the reliability of from their terminal into the system. The server receives this input information and tokenizes the text using the natural language processing library "spaCy". Specifically, it breaks down the information into words and phrases and extracts important elements such as nouns and verbs. This process identifies the subject and related words of the information. The input is information from the user, and the output is the extracted keywords.

[0127] Step 2:

[0128] The server uses a generative AI model to generate relevant keywords based on the keywords extracted in Step 1. These generated keywords are appropriately selected based on the content of the information. Specifically, the generative AI model understands the context of the information and adds highly relevant phrases. This process forms a foundation for a deeper understanding of the information's content. The input is the extracted keywords, and the output is the generated relevant keywords.

[0129] Step 3:

[0130] The server uses the keywords generated in step 2 to search external information sources. Specifically, it uses APIs such as "Google News API" and "Bing Search API" to retrieve relevant news articles and official announcements. This search collects data to verify the reliability of the information. The input is the generated related keywords, and the output is the search results obtained from the external information sources.

[0131] Step 4:

[0132] The server analyzes the search results obtained in step 3 and determines the reliability of the information. Specifically, it compares the search results and evaluates whether the information can be verified by other reliable sources. This evaluation determines the accuracy of the information and reports it to the user. The input is the search results from external sources, and the output is the evaluation result regarding the reliability of the information.

[0133] Step 5:

[0134] The server provides the user with the evaluation results obtained in step 4. Specifically, it notifies the user of its judgment on whether the information is reliable and, if necessary, provides detailed information sources. This notification allows the user to quickly and accurately determine the reliability of the information. The input is the evaluation result regarding the reliability of the information, and the output is the notification to the user.

[0135] (Application Example 2)

[0136] Next, we will describe application example 2 of form example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0137] In modern society, the internet contains a vast amount of news articles, some of which contain unreliable or misleading information. Making decisions based on such information can have significant consequences for individuals and society as a whole. Therefore, there is a need to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0138] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0139] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information resources on an information network, means for fact-checking, means for comparing with expert opinions, means for notifying the user of the evaluation results, means for analyzing the article content using a generative AI model, and means for generating prompt sentences and comparing them with the information sources. This makes it possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0140] A "news article" is a collection of information provided on the internet or in print, and includes reports on specific events or incidents.

[0141] "Reliability" is an indicator that shows that information is accurate and free from errors.

[0142] An "information network" is a system for acquiring, transmitting, and sharing information through the internet and other digital communication methods.

[0143] "Information resources" refer to databases, official announcements, and other reliable sources used to assess the credibility of news articles.

[0144] A "primary source" refers to an institution or organization that provides official announcements or data directly related to the events being reported.

[0145] "Fact-checking" is the process of verifying whether reported information is accurate by comparing it with other reliable sources.

[0146] "Expert opinion" refers to evaluations or comments provided by individuals or organizations with knowledge or experience in a particular field.

[0147] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to generate and analyze text.

[0148] A "prompt statement" is an instruction given to a generative AI model, containing instructions for performing a specific task.

[0149] "User" refers to an individual or organization that receives the results of a reliability assessment of a news article.

[0150] The system for implementing this invention uses a server and a user terminal to evaluate the reliability of news articles. The server analyzes the text of the news article and uses information resources on the information network to verify its match with the primary source. Specifically, the server analyzes the article content using a generative AI model, generates prompt sentences, and compares them with the source. This process makes it possible to evaluate the reliability of the article.

[0151] The server notifies the user terminal of the evaluation results. The user terminal receives the evaluation results and displays them to the user. This allows the user to quickly verify the reliability of the news article.

[0152] As a concrete example, consider a scenario where a user is reading a news article stating that "the number of people infected with the new virus is rapidly increasing." The server extracts information about the number of infected people from the article and compares it with official announcements from the Ministry of Health, Labour and Welfare and reports from other reliable media outlets. Based on the comparison results, the server evaluates the reliability of the article as "high," "medium," or "low" and notifies the user's terminal.

[0153] Examples of prompts to input into a generative AI model include the following:

[0154] "Please compare the content of the following news article with reliable sources and assess its reliability. Article content: Article text"

[0155] In this way, by using a server and user terminals, it is possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0156] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0157] Step 1:

[0158] A user views a news article. The user's device sends the text data of the news article being viewed to the server. The input is the text data of the news article, and the output is the transmission to the server.

[0159] Step 2:

[0160] The server analyzes the text data of the received news articles. Using a generative AI model, it extracts important facts and information from the articles. The input is the text data of the news articles, and the output is the extracted facts and information.

[0161] Step 3:

[0162] The server generates a prompt message based on the extracted facts and information. The generated prompt message includes instructions for cross-referencing with reliable sources. The input is the extracted facts and information, and the output is the generated prompt message.

[0163] Step 4:

[0164] The server uses the generated prompt message to search for information resources on the information network and verify a match with the primary source. The input is the generated prompt message, and the output is the matching result.

[0165] Step 5:

[0166] The server evaluates the reliability of news articles based on the matching results. The evaluation is categorized as "high," "medium," or "low." The input is the matching results, and the output is the reliability evaluation result.

[0167] Step 6:

[0168] The server notifies the user terminal of the evaluation results. The user terminal displays the received evaluation results to the user. The input is the reliability evaluation result, and the output is the notification to the user.

[0169] In this way, it is possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0170] (Example 3)

[0171] Next, we will describe Embodiment 3 of Embodiment Example 3. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0172] When evaluating the reliability of an informational article, it is necessary to efficiently check how well the article's content aligns with expert opinions and primary sources. However, traditional methods have the drawback of requiring recipients of information to spend a lot of time and effort judging the reliability of an article.

[0173] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 3 is realized by the following means.

[0174] In this invention, the server includes means for verifying the consistency of the information article text with primary sources using information from materials on an information network, means for verifying facts, and means for comparing with expert opinions. This makes it possible to quickly and accurately evaluate the reliability of the information article.

[0175] An "information article" is a piece of text, such as news or reports, provided on an information network.

[0176] "Reliability" refers to the characteristic that indicates information is accurate and free from errors and biases.

[0177] An "information network" is a system for acquiring, transmitting, and sharing information via the internet or other means.

[0178] A "primary source" is the origin of information and refers to documents or evidence that provide direct data or facts.

[0179] "Expert opinion" refers to views or comments provided by individuals with advanced knowledge and experience in a particular field.

[0180] "Natural language processing technology" refers to technologies that enable computers to understand, interpret, and generate human language.

[0181] A "data set" is a collection of data gathered for a specific purpose.

[0182] "Evaluation results" refer to information obtained as a result of evaluating the reliability and accuracy of the content of an informational article.

[0183] An "information terminal" is a device used for inputting, displaying, and processing information, and includes computers, smartphones, and other similar devices.

[0184] The following system configuration is used as an embodiment for carrying out this invention.

[0185] The server runs a program to evaluate the reliability of informational articles. This program analyzes the content of the informational articles using natural language processing techniques. Specifically, it uses a generative AI model to tokenize the content of the articles and extract important technical terms and claims. The server then compares the analyzed content of the informational articles with a data set containing expert opinions. This data set includes opinions from experts in various fields and is obtained from materials on the information network.

[0186] The terminal is responsible for sending informational article data entered by the user to the server. Users request analysis from the system by entering the URL or text of the article they want to analyze into the terminal. For example, if a user enters the prompt "I want to check opinions on the latest medical research," the terminal will send that information to the server.

[0187] The server generates and returns to the information terminal the results of its assessment of the reliability of the informational article. The assessment results include information about the article's reliability and its degree of agreement with expert opinions. Users can review these results through their terminal and determine the reliability of the informational article.

[0188] This system allows users to quickly and accurately verify how well the content of an informational article matches expert opinions and primary sources. The specific processing flow in Example 3 will be explained using Figure 15.

[0189] Step 1:

[0190] The user enters the URL or text of the information article they want to analyze into the terminal. The entered data is saved on the terminal as a prompt message. For example, the user might enter the prompt message, "I want to see opinions on the latest medical research."

[0191] Step 2:

[0192] The terminal sends the prompt text entered by the user to the server. The data sent includes the URL and text of the informational article, which serves as input data for the server to perform analysis.

[0193] Step 3:

[0194] The server analyzes the content of the informational article using natural language processing techniques based on the received prompt message. Specifically, it uses a generative AI model to tokenize the article's content and extract important technical terms and arguments. This process outputs the article's content as structured data.

[0195] Step 4:

[0196] The server compares the content of the analyzed informational article with a data set containing expert opinions. This data set includes opinions from experts in various fields and is obtained from materials on the information network. The degree of agreement between the article's claims and the expert opinions is then evaluated.

[0197] Step 5:

[0198] The server generates evaluation results and sends them back to the information terminal. These results include information about the article's reliability and its degree of agreement with expert opinions. This result becomes the output data from the server to the terminal.

[0199] Step 6:

[0200] The terminal displays the evaluation results received from the server to the user. Through the terminal, the user can check the reliability of the article and determine how well the content of the informational article matches expert opinions and primary sources.

[0201] (Application Example 3)

[0202] Next, we will describe application example 3 of form example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0203] While the internet contains a vast amount of news articles and information, it also includes misinformation and unreliable sources. This can lead readers and information recipients to make incorrect judgments, highlighting the need to efficiently evaluate the reliability of news articles and provide accurate information.

[0204] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means.

[0205] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, means for comparing with expert comments, means for analyzing the content of the article using natural language processing technology, and means for comparing with expert opinions and reliable sources using a generative AI model. This makes it possible to evaluate the reliability of news articles with high accuracy and provide accurate information to readers and information recipients.

[0206] A "news article" is a collection of information provided on the internet or in print media, and includes reports about specific events or incidents.

[0207] "Reliability" is an indicator that shows that information is accurate and free from errors.

[0208] A "primary source" refers to the original source or data from which information was first disseminated.

[0209] "Fact-checking" is the process of verifying whether the information provided is accurate.

[0210] "Expert comments" refer to opinions or views expressed by individuals with specialized knowledge in a particular field.

[0211] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[0212] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new information or data.

[0213] "Comparison" is the act of comparing two or more pieces of information or data and evaluating their similarities and differences.

[0214] The system for implementing this invention uses a server and a user terminal to evaluate the reliability of news articles. The server compares the text of the news article with literature information on the internet and verifies its consistency with primary sources. Furthermore, it performs fact-checking and cross-references it with expert comments. This makes it possible to evaluate the reliability of news articles with high accuracy.

[0215] The server analyzes the content of articles using natural language processing technology. Specifically, it uses a natural language processing library (e.g., spaCy) to extract keywords and arguments from the articles. Next, it uses a generative AI model (e.g., OpenAI's GPT-3) to compare the extracted information with expert opinions and reliable sources. This comparison is used to evaluate the reliability of the articles.

[0216] The user's terminal receives evaluation results provided by the server and displays them to the user. This allows the user to quickly check the reliability of an article and reduces the risk of being misled by misinformation.

[0217] As a concrete example, if a user is reading an article about the effectiveness of a new vaccine, the server analyzes the claims in the article and inputs a prompt message into an AI model asking, "How do medical experts evaluate the effectiveness of this vaccine?" The system then compares the expert opinions returned by the model with the content of the article and evaluates their reliability.

[0218] The flow of the specific processing in Application Example 3 will be explained using Figure 16.

[0219] Step 1:

[0220] A user views a news article. The user's device sends the text data of the news article being viewed to the server. The input is the text data of the news article, and the output is the transmission of data to the server.

[0221] Step 2:

[0222] The server analyzes the text data of received news articles using natural language processing techniques. Specifically, it uses a natural language processing library (e.g., spaCy) to extract keywords and main arguments from the articles. The input is the text data of the news articles, and the output is the extracted keywords and main arguments.

[0223] Step 3:

[0224] The server inputs prompt sentences into a generative AI model (e.g., OpenAI's GPT-3) based on the extracted keywords and claims. These prompt sentences are in the form of seeking expert opinions related to the article's content. The input consists of extracted keywords and claims, and the output is the generated prompt sentences.

[0225] Step 4:

[0226] The generative AI model generates expert opinions and information from reliable sources based on the input prompt sentence. The input is the prompt sentence, and the output is the generated expert opinion or information.

[0227] Step 5:

[0228] The server compares expert opinions and information obtained from the generated AI model with the content of the news article. The comparison evaluates the reliability of the article. The input is the content of the news article and the generated expert opinions, and the output is the reliability evaluation result.

[0229] Step 6:

[0230] The server sends the reliability evaluation results to the user terminal. The user terminal displays the received evaluation results to the user. The input is the reliability evaluation results, and the output is the display to the user.

[0231] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0232] "Example of form 1"

[0233] One embodiment of the present invention provides a news article reliability evaluation system incorporating an emotion engine. This system includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes an emotion engine that recognizes the user's emotions. Specifically, it estimates emotions from the user's facial expressions and tone of voice while reading the article, as well as keystrokes when typing text, and reflects the results in the reliability evaluation. For example, if the user shows anger or distrust, the system lowers the reliability of the article. Conversely, if the user shows satisfaction or trust, the system increases the reliability of the article.

[0234] "Example of form 2"

[0235] Another embodiment of the present invention provides a news article reliability evaluation system incorporating an emotion engine. This system includes means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes means for providing the evaluation results to the article's readers.

[0236] This also includes an emotion engine that recognizes the user's emotions. Specifically, the emotion engine recognizes the emotions the user feels when reading an article and adjusts the reliability rating based on those emotions. For example, if a user feels joy or excitement while reading an article, the article will be rated as highly reliable.

[0237] "Example of form 3"

[0238] As a further embodiment of the present invention, a news article reliability evaluation system incorporating an emotion engine is provided. This system includes means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes means for providing the evaluation results to the article writer and an emotion engine that recognizes the user's emotions. Specifically, the emotion engine recognizes the emotions the article writer felt when writing the article and adjusts the reliability evaluation based on those emotions. For example, if the writer felt anxiety or tension when writing the article, the reliability of that article will be evaluated as low.

[0239] The following describes the processing flow for each example of the form.

[0240] "Example of form 1"

[0241] Step 1: The user selects a news article.

[0242] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[0243] Step 3: The system verifies the facts.

[0244] Step 4: The system compares the information with expert comments.

[0245] Step 5: The emotion engine recognizes the user's emotions.

[0246] Step 6: The system incorporates the results of the emotion engine into the reliability assessment.

[0247] "Example of form 2"

[0248] Step 1: The user selects a news article.

[0249] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[0250] Step 3: The system verifies the facts.

[0251] Step 4: The system compares the information with expert comments.

[0252] Step 5: The emotion engine recognizes the user's emotions.

[0253] Step 6: The system incorporates the results of the emotion engine into the reliability assessment and provides the assessment results to the article's readers.

[0254] "Example of form 3"

[0255] Step 1: The writer creates the news article.

[0256] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[0257] Step 3: The system verifies the facts.

[0258] Step 4: The system compares the information with expert comments.

[0259] Step 5: The emotion engine recognizes the emotions of the article's writer.

[0260] Step 6: The system incorporates the results of the emotion engine into the reliability assessment and provides the assessment results to the article writer.

[0261] (Example 1)

[0262] Next, we will describe Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0263] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, with the vast amount of information available on the internet, verifying whether an article's content aligns with primary sources is not easy. Furthermore, while considering readers' sentiments and expert opinions is necessary when evaluating article reliability, there is a lack of efficient means to do so.

[0264] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0265] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, and means for recognizing user sentiment and reflecting the results in reliability evaluation. This makes it possible to evaluate the reliability of news articles from multiple perspectives and provide accurate evaluation results.

[0266] A "news article" refers to written text or news content created to convey information, and is particularly widely distributed through the internet and print media.

[0267] "Reliability" is an indicator that shows that information or data is accurate and error-free, and it represents the degree to which the recipient of the information can trust its content.

[0268] An "information network" refers to a communication infrastructure for sending and receiving digital data, including the internet, and is a system that allows access to various information sources.

[0269] A "primary source" refers to the place where specific information or data first originated, or the entity that directly provides that information, and forms the basis of reliable information.

[0270] "Fact-checking" is the process of verifying whether the information or data provided matches the actual facts, and is carried out to guarantee the accuracy of the information.

[0271] "Expert opinions" refer to views and analyses provided by individuals or organizations with advanced knowledge and experience in a particular field, and serve as a reference when evaluating the reliability of information.

[0272] "User emotions" refer to the psychological reactions and feelings that users exhibit when reading news articles, and are inferred from their facial expressions, tone of voice, input behavior, etc.

[0273] "Reliability assessment" is an evaluation process used to determine how accurate and reliable the content of a news article is, and it is carried out by considering various factors.

[0274] This invention is a system for evaluating the reliability of news articles, in which the server, terminal, and user elements work together.

[0275] When the server receives a news article, it analyzes the article text using natural language processing techniques. Specifically, the server tokenizes the text and extracts important keywords and numerical data. Common natural language processing libraries can be used for this analysis. Next, the server uses databases on the information network to compare the extracted data with primary sources. Databases such as Google Scholar and PubMed are used to search for information that matches the article's content and calculate the degree of match.

[0276] The device senses the user's facial expressions, tone of voice, and keystrokes when they are typing text while reading news articles. This uses hardware such as a camera, microphone, and keyboard. The device collects this data in real time and estimates the user's emotions. The emotion engine analyzes this emotion data and sends the emotions the user is expressing to the server.

[0277] The server adjusts the reliability evaluation of a news article based on emotion data received from a terminal. When the user expresses anger or distrust, the server evaluates the reliability of the article as low. Conversely, when the user expresses satisfaction or trust, the server evaluates the reliability of the article as high.

[0278] As a specific example, consider a case where a user is reading a "news article regarding the efficacy of a new virus vaccine". The server searches for the data "95% vaccine efficacy" mentioned in the article on an information network, and confirms whether it matches the primary information source. At the same time, the terminal detects the user's facial expression and voice tone, and if the user expresses distrust, the server evaluates the reliability of the article as low.

[0279] An example of a prompt sentence input to a generative AI model is: "To evaluate the reliability of a news article, please explain a method for adjusting reliability by collating data in the article with primary information sources on an information network and taking user emotion data into consideration". Through this prompt sentence, the generative AI model can generate a text that explains the processing of the system in detail.

[0280] The flow of specific processing in Embodiment 1 will be described with reference to FIG. 17.

[0281] Step 1:

[0282] The server receives a news article from a user. The body of the news article is provided as an input. The server analyzes the article body using natural language processing technology, and extracts important keywords and numerical data. Specifically, the text is tokenized, and parts of speech such as nouns and verbs are identified. A list of the extracted keywords and data is generated as an output.

[0283] Step 2:

[0284] The server searches databases on the information network based on the data extracted in Step 1. Keywords and a list of data are used as input. The server uses databases such as Google Scholar and PubMed to find information that matches the primary sources. Specifically, it sends queries to the databases via APIs to retrieve relevant literature and data. A score indicating the degree of matching is generated as output.

[0285] Step 3:

[0286] The device collects user emotion data when the user reads news articles. Input includes the user's facial expressions, voice tone, and keystrokes during text input. The device uses its camera, microphone, and keyboard to sense and collect this data in real time. Specifically, it uses facial recognition and voice analysis technologies to estimate the user's emotions. The estimated emotion data is then generated as output.

[0287] Step 4:

[0288] The server integrates the similarity score obtained in step 2 with the sentiment data obtained in step 3 to evaluate the reliability of the news article. The similarity score and sentiment data are used as input. The server uses a sentiment engine to calculate the impact of the user's emotions on the reliability evaluation. Specifically, if anger or distrust is indicated, the reliability is rated low, and if satisfaction or trust is indicated, the reliability is rated high. The final reliability evaluation score is generated as output.

[0289] Step 5:

[0290] The server presents the final reliability rating score to the user. The reliability rating score is used as input. The server displays the evaluation results visually, making them easy for the user to understand. Specifically, the reliability score is shown numerically and graphically, providing the user with information to judge the reliability of the article. A visualized evaluation result is generated as output.

[0291] (Application Example 1)

[0292] Next, we will describe Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server," and the smart device 14 will be referred to as a "terminal."

[0293] In modern society, where the reliability of information is paramount, there is a need to quickly and accurately evaluate the reliability of news articles. However, conventional methods are time-consuming in verifying the accuracy of primary sources and fact-checking, and they do not take into account the emotions of users in their reliability assessments. As a result, it is difficult to provide information that is truly reliable to recipients.

[0294] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0295] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, and means for recognizing user sentiment and reflecting the results in the reliability evaluation. This enables rapid and accurate evaluation of the reliability of news articles and personalized reliability evaluations that take user sentiment into consideration.

[0296] A "news article" is a piece of writing intended to convey information, and in particular, it includes content related to current events and social issues.

[0297] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[0298] An "information network" refers to a communication system for sending and receiving digital data, including the internet.

[0299] The term "primary information source" refers to materials and data that provide the origin of information or direct evidence.

[0300] The term "fact-checking" refers to the process for verifying whether information is accurate.

[0301] The term "expert opinion" refers to the views provided by a person with advanced knowledge and experience in a specific field.

[0302] The term "user" refers to an individual or organization that uses a system or service.

[0303] The term "emotion recognition" refers to the process of estimating a user's emotional state from the user's facial expression, voice tone, behavior, and the like.

[0304] The term "reliability evaluation" refers to the process of indicating the accuracy and reliability of information using numerical values or indicators.

[0305] The term "personalization" refers to adjusting services or information in accordance with the characteristics and preferences of individual users.

[0306] In the system for carrying out the present invention, a server and a terminal operate in cooperation with each other to evaluate the reliability of a news article. The server receives the main text of the news article, and uses literature information on an information network to check for consistency with a primary information source. In this way, fact-checking of the article is performed, and cross-checking with expert opinions is implemented.

[0307] When a user browses a news article, the terminal acquires the user's facial expression and voice tone using a camera and a microphone. In this way, emotion recognition is performed, and the result is transmitted to the server. The server adjusts the reliability evaluation based on the received emotion data, and calculates a final reliability score.

[0308] This system will be implemented as an application installed on devices such as smartphones and smart glasses. Specifically, it will perform natural language processing and sentiment analysis using software such as Python and TensorFlow. spaCy will be used as the natural language processing library, and NLTK will be used as the sentiment analysis library.

[0309] For example, when a user is reading a news article through smart glasses, the article's reliability score is displayed in their field of vision. If the user frowns, the emotion engine detects the distrust and adjusts the reliability score.

[0310] An example of a prompt to input into a generative AI model is: "Please rate the reliability of this news article. The article content is as follows: 'Article Content'. The user's sentiment indicates distrust."

[0311] The flow of a specific process in Application Example 1 will be explained using Figure 18.

[0312] Step 1:

[0313] The device uses its camera and microphone to capture the user's facial expressions and voice tone when they view news articles. This data is used as input for sentiment analysis.

[0314] Step 2:

[0315] The device inputs the acquired facial and voice data into an emotion analysis library (such as NLTK) to estimate the user's emotions. The result of the emotion analysis is output, indicating the emotion the user is expressing (e.g., distrust, satisfaction).

[0316] Step 3:

[0317] The terminal sends the text of the news article to the server. The server analyzes the received article using a natural language processing library (such as spaCy) and extracts facts and data from the article.

[0318] Step 4:

[0319] The server compares the extracted facts and data with bibliographic information on the information network and calculates the degree of match with the primary source. This degree of match is output as the basis for reliability evaluation.

[0320] Step 5:

[0321] The server compares the article content with an expert opinion database and evaluates the degree of agreement between the expert's view and the article. This evaluation result is also taken into account in the reliability assessment.

[0322] Step 6:

[0323] The server receives user sentiment data and incorporates it into the reliability assessment. Specifically, if a user expresses distrust, the reliability score is adjusted.

[0324] Step 7:

[0325] The server calculates the final reliability score and sends it to the terminal. The terminal then displays this score visually to the user.

[0326] Step 8:

[0327] Users can check the displayed reliability score to determine the reliability of a news article. This allows users to read articles while considering the accuracy of the information.

[0328] (Example 2)

[0329] Next, we will describe Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0330] When evaluating the reliability of news articles, there is a challenge in objectively judging the accuracy and reliability of the information. Furthermore, since the influence of user emotions on the reliability evaluation of an article is not taken into account, the evaluation results may be influenced by the user's subjectivity. In addition, there is a lack of means to visually present the evaluation results, making it difficult for users to intuitively understand the results.

[0331] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0332] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, means for recognizing user sentiment and adjusting reliability evaluation based on that sentiment, and means for visually displaying the evaluation results. This enables objective and intuitive evaluation of the reliability of news articles, and allows for reliability evaluation that takes user sentiment into consideration.

[0333] A "news article" is written or digital content created to report events or information.

[0334] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[0335] An "information network" is a system for acquiring, transmitting, and sharing information through the internet and other digital communication methods.

[0336] A "primary source" refers to the direct origin or source of an event or piece of information.

[0337] "Fact-checking" is the process of verifying whether information is accurate using other reliable sources.

[0338] "Expert opinion" refers to the views of an individual or organization that possesses advanced knowledge and experience in a particular field.

[0339] "User sentiment" refers to the psychological reactions that users exhibit when reading news articles.

[0340] "Visual display methods" refer to methods of presenting information to users in an easy-to-understand manner using graphs, diagrams, numerical data, etc.

[0341] As an embodiment of this invention, a system for evaluating the reliability of news articles is constructed. The server receives news articles and analyzes their content using natural language processing technology. Specifically, it uses a text analysis library (e.g., spaCy or NLTK) to extract important facts and keywords from the articles.

[0342] The server searches for reliable sources of information through its information network based on extracted facts and keywords. Using APIs, it accesses official police announcements and other media databases to verify the facts presented in the articles. It also sends queries to external databases using HTTP requests.

[0343] The device recognizes the user's emotions as they read the article. It uses the camera and microphone to capture the user's facial expressions and voice, which are then analyzed by an emotion engine. The emotion engine uses machine learning models (e.g., TensorFlow or PyTorch) to classify the user's emotions into categories such as "joy," "anxiety," and "excitement."

[0344] The server integrates fact-checking results with user sentiment assessment results to evaluate the reliability of news articles. If a user expresses positive emotions after reading an article, the reliability score increases. Conversely, if a user expresses negative emotions, more emphasis is placed on cross-referencing with other sources.

[0345] Users receive the reliability evaluation results through their device. The device visually displays the evaluation results, providing the article's reliability score and comments. The evaluation results are presented in graphs and numerical values, in a way that is easy for users to understand.

[0346] As a concrete example, when a user enters a news article about the spread of a new virus, the server analyzes the article's content and verifies relevant facts using external sources. The device captures the user's facial expressions with its camera and analyzes them with an emotion engine. Finally, the reliability evaluation results are displayed on the device, allowing the user to verify the article's reliability.

[0347] Examples of prompts for a generative AI model:

[0348] "Analyze news articles about the spread of the new virus and evaluate their reliability. Verify the facts in the articles with other reliable sources and adjust your evaluation considering user sentiment."

[0349] The flow of the specific processing in Example 2 will be explained using Figure 19.

[0350] Step 1:

[0351] The server receives news articles from users as input. The server analyzes the article content using natural language processing techniques, extracting important facts and keywords. This process utilizes text analysis libraries (e.g., spaCy or NLTK) to analyze the article's grammatical structure and identify keywords. The server then generates a list of extracted keywords and facts as output.

[0352] Step 2:

[0353] The server uses the keywords and facts extracted in Step 1 as input to search for reliable sources of information through the information network. The server accesses external databases using APIs to verify whether the facts in the articles can be confirmed. Specifically, it sends queries to external databases using HTTP requests and retrieves the matching results as output.

[0354] Step 3:

[0355] The device recognizes the user's emotions as input while reading news articles. The device uses its camera and microphone to capture the user's facial expressions and voice, which are then analyzed by an emotion engine. The emotion engine uses machine learning models (e.g., TensorFlow or PyTorch) to classify the user's emotions into categories such as "joy," "anxiety," and "excitement." As output, it generates user emotion data.

[0356] Step 4:

[0357] The server integrates the matching results from Step 2 and the sentiment data from Step 3 as input to evaluate the reliability of the news article. If the user expresses positive sentiment after reading the article, the server increases the reliability score. Conversely, if the user expresses negative sentiment, it places more emphasis on the matching results with other sources. The server generates a reliability evaluation score as output.

[0358] Step 5:

[0359] The user receives the reliability evaluation results through their device. The device visually displays the evaluation results and provides the article's reliability score and comments. Specifically, it presents the evaluation results in graphs and numerical values ​​in a way that is easy for the user to understand. As output, it generates a visually displayed reliability evaluation result.

[0360] (Application Example 2)

[0361] Next, we will describe application example 2 of form example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0362] When evaluating the reliability of news articles, it is necessary to consider not only the content but also the reader's emotions. However, conventional systems do not perform reliability evaluations that take emotions into account, which presents a challenge in providing readers with reliable information.

[0363] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0364] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, and means for recognizing the user's emotions and adjusting the reliability evaluation based on those emotions. This makes it possible to evaluate the reliability of news articles while taking the reader's emotions into consideration and to provide highly reliable information.

[0365] A "news article" refers to written text or news reports created to convey information.

[0366] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[0367] An "information network" refers to a communication network used to transmit information, including the internet.

[0368] A "primary source" is a source that provides the origin of information or direct evidence.

[0369] "Fact-checking" is the process of verifying whether information is accurate.

[0370] "Expert comments" refer to opinions or explanations provided by individuals with specialized knowledge in a particular field.

[0371] "Users" refer to individuals or organizations that use a system or service.

[0372] "Emotion" refers to a person's psychological reaction or state, and includes feelings such as joy and sadness.

[0373] "Evaluation" refers to the act of judging value or performance based on specific criteria.

[0374] A "server" is a computer system that provides information and services over a network.

[0375] The system for carrying out this invention includes a program for evaluating the reliability of news articles. The server analyzes the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. This is done using natural language processing libraries (e.g., NLTK, spaCy). Furthermore, it uses external APIs (e.g., Google Fact Check Tools API) to verify the accuracy of the article for fact-checking.

[0376] The device acquires data through its camera and microphone to recognize the user's emotions and analyzes it using an emotion recognition library (e.g., Affectiva). This allows for adjusting the reliability evaluation based on the user's emotions. The evaluation results are displayed on the device's screen and provided to the user.

[0377] As a concrete example, when a user opens a news app on their smartphone and begins reading an article, the server evaluates the article's reliability and displays a reliability score on the screen. If the user smiles while reading the article, the device recognizes that emotion and adjusts the reliability score accordingly.

[0378] An example of a prompt to input into a generative AI model is: "Please evaluate the reliability of this news article. Compare the article's content with external sources and calculate a reliability score, taking into account the user's sentiment."

[0379] The flow of a specific process in Application Example 2 will be explained using Figure 20.

[0380] Step 1:

[0381] The server retrieves the text of the news article selected by the user. It receives the URL or ID of the news article as input and extracts the text data of the article. It generates the text data of the news article to be analyzed as output.

[0382] Step 2:

[0383] The server analyzes the text data of news articles using natural language processing libraries (e.g., NLTK, spaCy). It receives the text data of news articles as input and extracts keywords and important phrases from the articles. As output, it generates data that summarizes the content of the articles.

[0384] Step 3:

[0385] The server uses external APIs (e.g., Google Fact Check Tools API) to fact-check the content of news articles. It receives summary data as input and compares it with external sources. As output, it generates initial assessment data regarding the reliability of the article.

[0386] Step 4:

[0387] The device acquires data on the user's facial expressions and voice through its camera and microphone in order to recognize the user's emotions. It receives real-time video and audio data of the user as input and analyzes it using an emotion recognition library (e.g., Affectiva). As output, it generates data indicating the user's emotional state.

[0388] Step 5:

[0389] The server adjusts the reliability rating of news articles based on the user's emotional state data. It receives initial evaluation data and emotional state data as input and recalculates the reliability score. It generates the adjusted reliability score as output.

[0390] Step 6:

[0391] The device displays the adjusted reliability score to the user. It receives the adjusted reliability score as input and displays it visually on the user's screen. As output, it provides a reliability evaluation result that the user can visually confirm.

[0392] (Example 3)

[0393] Next, we will describe Embodiment 3 of Embodiment Example 3. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0394] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, verifying whether an article's content aligns with primary sources and expert opinions is not easy. Furthermore, it is necessary to consider the impact of the writer's emotions on the article's reliability. An effective system is needed to address these challenges.

[0395] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 3 is realized by the following means.

[0396] In this invention, the server includes means for analyzing information, means for confirming matching with information sources, and means for verifying facts. This makes it possible to evaluate the reliability of news articles from multiple perspectives and to quickly and accurately determine their reliability.

[0397] "Means of analyzing information" refers to a function that uses natural language processing technology to analyze the content of news articles and extract expert arguments and keywords.

[0398] "Means of verifying consistency with information sources" refers to the function of searching relevant literature and databases and comparing their contents to confirm whether the claims in a news article are consistent with primary sources.

[0399] "Means of fact-checking" refers to the function of verifying the accuracy of information contained in news articles by referring to relevant data and expert opinions.

[0400] "Means of comparing with expert opinions" refers to a function that compares and evaluates the claims made in news articles with the opinions and comments of experts in that field.

[0401] "Means of recognizing emotions" refers to the function of analyzing the emotions of news article writers and evaluating the impact of those emotions on the article's credibility.

[0402] "Means for adjusting evaluations" refers to a function that adjusts the reliability evaluation of news articles based on collected information and sentiment analysis results, and generates a final evaluation result.

[0403] To implement this invention, it is necessary to build a system for evaluating the reliability of news articles. The user inputs the news article they wish to evaluate into the system. The server analyzes the content of the article using natural language processing techniques and extracts expert claims and keywords. This analysis uses the Python natural language processing library.

[0404] Next, the server gathers relevant literature and expert opinions from the internet based on the extracted claims. Specifically, it uses academic paper database APIs to search for relevant literature and compare it with the claims in the article. Furthermore, the server uses database query techniques to verify that the claims in the article match those of primary sources.

[0405] Furthermore, the server uses an emotion engine to analyze the emotions of the article's writer. Specifically, it uses a Python emotion analysis library to determine whether the writer is feeling anxious or stressed. This emotion information influences the reliability evaluation.

[0406] Finally, the server integrates the collected information and sentiment analysis results to evaluate the reliability of the news article. The evaluation results are displayed visually to the user. For example, the reliability score may be shown in a graph, with a detailed explanation of which factors influenced the evaluation.

[0407] As a concrete example, when a user enters "an article about the effectiveness of a new virus vaccine," the server analyzes the article and extracts claims about vaccine effectiveness. Next, the server searches for relevant medical papers using an academic paper database and evaluates the accuracy of the claims. Simultaneously, an emotion engine analyzes the writer's emotions and incorporates this into the reliability assessment. Finally, the server generates a reliability score and provides it to the user.

[0408] An example of a prompt to the generative AI model might be: "Please verify whether the medical claims in this article are consistent with expert opinions. Also, please evaluate the impact of the writer's emotions on its credibility." The specific processing flow in Example 3 will be explained using Figure 21.

[0409] Step 1:

[0410] The user inputs a news article they want to evaluate into the system. The server receives the input article and analyzes its content using natural language processing techniques. Specifically, it uses a Python natural language processing library to extract expert claims and keywords from the article. As a result of this analysis, a list of the article's claims and keywords is output.

[0411] Step 2:

[0412] Based on the claims extracted in Step 1, the server collects relevant literature and expert opinions from the internet. Specifically, it uses an academic paper database API to search for relevant literature and compares it with the claims in the article. This process takes a list of claims as input and produces a list of relevant literature and expert comments as output.

[0413] Step 3:

[0414] The server verifies whether the claims in the article match those of the primary source. Specifically, it uses database query techniques to identify the primary source and check for content consistency. This step takes a list of claims as input and outputs results indicating whether or not there is a match.

[0415] Step 4:

[0416] The server uses an emotion engine to analyze the emotions of the article's author. Specifically, it uses a Python emotion analysis library to determine whether the author is feeling anxious or stressed. This process takes the article's text as input and outputs the results of the emotion analysis.

[0417] Step 5:

[0418] The server integrates the collected information and sentiment analysis results to evaluate the reliability of the news article. Specifically, it calculates a reliability score and generates an evaluation result. In this step, relevant literature, expert comments, agreement results, and sentiment analysis results are used as input, and the reliability score is obtained as output.

[0419] Step 6:

[0420] The server provides the user with the final evaluation results. Specifically, it visually displays the reliability score and provides a detailed explanation of which factors influenced the evaluation. In this step, the reliability score is used as input, and the evaluation results are displayed to the user as output.

[0421] (Application Example 3)

[0422] Next, we will describe application example 3 of form example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0423] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, verifying whether an article's content aligns with expert opinions and primary sources is not easy, nor is it feasible to consider the impact of the writer's emotions on its reliability. Therefore, there is a need for a system that comprehensively evaluates the reliability of news articles and provides this information to readers and writers.

[0424] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means.

[0425] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, means for analyzing the sentiment of the article's writer and adjusting the reliability evaluation, and means for displaying the evaluation results on a display device. This makes it possible to comprehensively evaluate the reliability of news articles and provide this information to readers and writers quickly and accurately.

[0426] A "news article" is a piece of writing created to convey information, and includes reporting on specific events or incidents.

[0427] "Reliability" is a concept that refers to the degree to which information is judged to be accurate and free from errors.

[0428] An "information network" is a system that uses digital communication technologies such as the internet to share information.

[0429] A "primary source" refers to the place or person from which specific information or data first originated; it is the original source of the information.

[0430] "Expert opinion" refers to views or comments from individuals who possess advanced knowledge and experience in a particular field.

[0431] "Sentiment analysis" is a technique that extracts emotions from written text or speech and evaluates the type and intensity of those emotions.

[0432] "Evaluation results" refer to the results of an evaluation based on specific criteria, and in this context, it means an evaluation of the reliability of a news article.

[0433] A "display device" is a device used to visually display information, and includes computer monitors and smartphone screens.

[0434] An "information terminal" is an electronic device used to process and display information, and includes smartphones and tablets.

[0435] The term "creator" refers to the person who produced a particular piece of writing or work.

[0436] The system for implementing this invention uses a server and an information terminal to evaluate the reliability of news articles. The server analyzes the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. Furthermore, it performs fact-checking and compares it with expert opinions. This makes it possible to comprehensively evaluate the reliability of the article.

[0437] The server uses a sentiment analysis API (e.g., IBM Watson® Natural Language Understanding) to analyze the sentiment of the article's author and adjust the reliability rating accordingly. The evaluation results are displayed on the information terminal's display device, allowing users to check them in real time.

[0438] As a concrete example, when a user opens a news app on their smartphone and selects a specific article, the article's reliability score is displayed on the screen. For instance, if the article reports on a "new medical discovery," the system checks whether the content aligns with the opinions of medical experts and calculates a reliability score.

[0439] An example of a prompt for a generative AI model is, "Verify whether the medical claims in this article are consistent with expert opinions and calculate a reliability score." This prompt prompts the server to perform the necessary data processing and reliability assessment.

[0440] The flow of the specific processing in Application Example 3 will be explained using Figure 22.

[0441] Step 1:

[0442] The user launches a news app on their information terminal and selects a specific article. The input is the news article selected by the user, and the output is the text data of that article. The terminal sends this text data to the server.

[0443] Step 2:

[0444] The server analyzes the text data of received news articles and uses bibliographic information on the information network to verify their match with primary sources. The input is the text data of news articles, and the output is the result of matching with primary sources. The server analyzes the text using a natural language processing library (e.g., spaCy) and verifies the match by calling an external fact-checking API.

[0445] Step 3:

[0446] The server verifies the content of news articles and compares them with expert opinions. The input is the text data of the news article, and the output is the results of the fact-checking and the agreement with expert opinions. The server refers to an expert database to check whether the content of the article is consistent with the opinions of experts.

[0447] Step 4:

[0448] The server uses a sentiment analysis API to analyze the writer's sentiment and adjust the reliability rating accordingly. The input is the text data of the news article, and the output is the result of the writer's sentiment analysis. The server calls a sentiment analysis API (e.g., IBM Watson Natural Language Understanding) to evaluate the writer's sentiment.

[0449] Step 5:

[0450] The server integrates these results to calculate the reliability score of the news article. The inputs are the results of matching primary sources, fact-checking, agreement with expert opinions, and sentiment analysis, and the output is the reliability score. The server integrates this data and runs an algorithm to calculate the reliability score.

[0451] Step 6:

[0452] The server sends the calculated reliability score to the information terminal, and the terminal displays the score to the user. The input is the reliability score, and the output is the reliability score displayed on the terminal's display device. The terminal displays the received score on the screen so that the user can verify it.

[0453] (Other examples)

[0454] Next, other embodiments will be described. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0455] In recent years, a vast number of news articles exist on the internet, and there is a need to quickly and accurately evaluate their reliability. However, traditional methods require a great deal of time and effort to determine the reliability of an article, making efficient evaluation difficult.

[0456] The identification process performed by the identification processing unit 290 of the data processing device 12 in other embodiments is realized by the following means.

[0457] In this invention, the server includes means for acquiring the text of a news article, means for generating prompts to search for bibliographic information on an information network based on the acquired news article text and to confirm the match with the primary source, and means for inputting the generated prompts into a generating AI model and evaluating the degree of match with the primary source. This makes it possible to quickly and accurately evaluate the reliability of a news article.

[0458] A "news article" is a piece of text containing information about current events or social occurrences, distributed via the internet or other media.

[0459] "Reliability" is an indicator that shows that information is accurate and free from errors and biases.

[0460] An "information network" is a collection of information resources connected through the internet and other digital communication methods.

[0461] A "primary source" is the initial source of specific information or data, and is considered the most reliable source of information.

[0462] A "prompt" is an input sentence used to give specific instructions or questions to a generative AI model.

[0463] A "generative AI model" is an algorithm that uses artificial intelligence technology to generate new information or data based on given input.

[0464] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[0465] "Evaluation results" refer to the conclusions or scores obtained after evaluating the reliability of a news article.

[0466] ---

[0467] This invention is a system for evaluating the reliability of news articles, and is realized through the respective roles of the server, terminal, and user.

[0468] The server first retrieves the news article text from the database to evaluate the reliability of the news article. This database is managed using a relational database management system such as MySQL®. The server then executes SQL queries to retrieve the necessary article data.

[0469] Next, the server generates a prompt message for input to the generation AI model based on the retrieved news article text. This prompt message includes instructions to check which primary source the news article's content matches. For example, the prompt message might be in the format: "Please check which primary source the content of this news article matches." The server uses a Python script to create a summary of the article and incorporates it into the prompt message.

[0470] The generated prompt text is input to a generative AI model such as OpenAI's GPT-4®. The server sends the prompt text to the generative AI model, which is hosted on the cloud, via an API and waits for a response from the model. This response includes the degree of match with the primary source.

[0471] Furthermore, the server searches for and verifies facts and data within news articles from publicly available sources on the information network. This process uses search engines such as Elasticsearch®. The server generates search queries, searches public sources, and retrieves results.

[0472] The server calculates a reliability score for news articles based on matching information from the generative AI model and search results from publicly available sources. This calculation uses NLTK, a Python natural language processing library, and applies a scoring algorithm.

[0473] Finally, the server sends the calculated reliability score to the user's device. The device then displays the evaluation results visually using React.js. The user can then check the evaluation results on their device and judge the reliability of the news article.

[0474] This system allows users to quickly and accurately assess the reliability of news articles.

[0475] The flow of specific processing in other embodiments will be explained using Figure 23.

[0476] Step 1:

[0477] To evaluate the reliability of news articles, the server first retrieves the article text from the database. Given an article ID as input, the server uses MySQL to execute an SQL query and retrieves the corresponding news article text as output. This process prepares the article data to be evaluated.

[0478] Step 2:

[0479] The server generates prompts for input to the generative AI model based on the retrieved news article text. Given the news article text as input, the server uses a Python script to create a summary of the article and outputs a prompt that reads, "Please check which primary sources the content of this news article matches." This prompt contains specific instructions for the generative AI model.

[0480] Step 3:

[0481] The server inputs the generated prompt text into a generative AI model. Given the prompt text as input, the server sends it to a generative AI model, such as OpenAI's GPT-4, via an API. The model returns a degree of matching with the primary source as output. This process provides an initial assessment of the article's reliability.

[0482] Step 4:

[0483] The server searches for and compares facts and data from news articles with publicly available information sources on the information network. Given the text of a news article as input, the server uses Elasticsearch to generate search queries and search public sources. The output is the search results. This process allows verification of how well the article's content matches information from other sources.

[0484] Step 5:

[0485] The server calculates a reliability score for news articles based on matching information from a generative AI model and search results from publicly available sources. Given matching information and search results as input, the server applies a scoring algorithm using NLTK, a Python natural language processing library, and calculates a reliability score as output. This process quantifies the overall reliability of the article.

[0486] Step 6:

[0487] The server sends the calculated reliability score to the user's device. The reliability score is provided as input, and the server sends the score to the user's device using the HTTP protocol. The device then visually displays the evaluation results using React.js. As output, the user can check the evaluation results on their device and judge the reliability of the news article. This process allows the user to quickly evaluate the reliability of an article.

[0488] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0489] Data generation model 58 is a form of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0490] Other examples of generative AI include Gemini® (registered trademark) (Internet search). <url: https: gemini.google.com ?hl="ja">) are examples.

[0491] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0492] [Second Embodiment]

[0493] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0494] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0495] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0496] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0497] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0498] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0499] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0500] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0501] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0502] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0503] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0504] Next, the identification process performed by the identification processing unit 290 of the data processing device 12 will be described.

[0505] "Example of form 1"

[0506] One embodiment of the present invention is a system for evaluating the reliability of news articles. This system has a means for verifying the consistency between the text of a news article and its primary source using information from online sources. Specifically, it compares the facts and data described in the article with publicly available papers and databases on the internet and calculates the degree of agreement.

[0507] "Example of form 2"

[0508] Furthermore, the system of the present invention also includes means for fact-checking. This verifies whether the facts and information described in the article can be confirmed by other reliable sources. For example, it verifies whether the incident reported in the article has been reported in official police statements or by other media outlets.

[0509] "Example of form 3"

[0510] Furthermore, the system of the present invention includes means for cross-referencing with expert comments. This means that the specialized content and views described in the article are evaluated in comparison with the opinions and comments of experts in that field. For example, it checks whether the medical claims reported in the article are consistent with comments and papers from experts in the medical community.

[0511] The following describes the processing flow for each example of the form.

[0512] "Example of form 1"

[0513] Step 1: The system receives the news article text as input.

[0514] Step 2: Extract the facts and data contained in the article.

[0515] Step 3: Compare the extracted facts and data with publicly available papers and databases on the internet and calculate the degree of agreement.

[0516] "Example of form 2"

[0517] Step 1: The system receives the news article text as input.

[0518] Step 2: Extract the facts and information contained in the article.

[0519] Step 3: Verify whether the extracted facts and information can be confirmed by other reliable sources.

[0520] "Example of form 3"

[0521] Step 1: The system receives the news article text as input.

[0522] Step 2: Extract the specialized content and viewpoints mentioned in the article.

[0523] Step 3: Evaluate the extracted specialized content and views by comparing them with the opinions and comments of experts in that field.

[0524] (Example 1)

[0525] Next, we will describe Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0526] There is a need to quickly and accurately assess the reliability of news articles, but traditional methods are time-consuming to cross-reference and evaluate information, making it difficult to obtain reliable results.

[0527] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0528] In this invention, the server includes means for verifying the match between the news article text and primary sources using information from literature on an information network, means for analyzing the article content using natural language processing technology, and means for searching publicly available sources on an information network and comparing facts and data within the article. This makes it possible to quickly and accurately evaluate the reliability of news articles.

[0529] A "news article" is a piece of writing intended to convey information, providing details about a specific event or occurrence.

[0530] "Reliability" is an indicator that shows that information is accurate and error-free, and it is a standard for evaluating the legitimacy and credibility of information.

[0531] An "information network" is a communication infrastructure for sending and receiving digital data, including the internet.

[0532] A "primary source" is the initial source of specific information or data, and is a source that provides direct evidence or data.

[0533] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and it involves analyzing text and extracting meaning.

[0534] "Public information sources" are databases and documents that provide information that is generally accessible to the public, and are collections of information that anyone can use.

[0535] A "text similarity calculation algorithm" is a computational method for numerically evaluating the similarity between different texts, and it measures the degree of textual agreement.

[0536] "Evaluation results" refer to the outcome of an evaluation conducted based on specific criteria, and indicate conclusions regarding reliability and accuracy.

[0537] To implement this invention, it is necessary to build a system for evaluating the reliability of news articles. The user sends the text of a news article to the server. The server analyzes the received news article using natural language processing technology. This analysis utilizes natural language processing libraries such as "spaCy" and "NLTK". The server extracts facts and data from the article and searches publicly available information sources on the information network. This search utilizes academic paper search APIs and database APIs.

[0538] The server compares the content of news articles with primary sources obtained through searches. For this comparison, it uses text similarity calculation algorithms such as "Cosine Similarity" and "Jaccard Index." The server uses these algorithms to calculate the degree of similarity and generate evaluation results. These evaluation results are provided to the user via a web interface or API.

[0539] As a concrete example, consider a case where a user wants to evaluate a news article about "the global economic growth rate in 2023." The server extracts the statement "the global economic growth rate in 2023 is 3.5%" from the article. Next, it uses an academic paper search API to search for related papers using the keyword "2023 global economic growth rate." The server compares the content of the papers obtained from the search results with the description in the article and calculates the degree of similarity using Cosine Similarity. Finally, the server evaluates the reliability as higher if the degree of similarity is high and returns the result to the user.

[0540] An example of a prompt to input into a generative AI model might be, "To evaluate the reliability of the news article, please compare the facts in the article with online sources."

[0541] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0542] Step 1:

[0543] The user sends the body of a news article to the server. The text data of the news article is provided as input. The server receives this text data and prepares for the next analysis step.

[0544] Step 2:

[0545] The server analyzes the text of received news articles using natural language processing techniques. Specifically, it uses the natural language processing library "spaCy" to extract facts and data from the articles. The input is the text data of the news article, and the output is a list of the extracted facts and data.

[0546] Step 3:

[0547] The server searches publicly available information sources on the information network based on the extracted facts and data. This search uses an academic paper search API. The input is a list of extracted facts and data, and the output is a list of relevant primary sources. Specifically, the server performs a search using the keyword "2023 global economic growth rate".

[0548] Step 4:

[0549] The server compares the content of news articles with primary sources obtained through searches. The "Cosine Similarity" algorithm, a text similarity calculation algorithm, is used for this comparison. The input consists of the text data of the news articles and a list of primary sources; the output is a numerical score indicating the degree of similarity. The server uses this to evaluate the reliability of the articles.

[0550] Step 5:

[0551] The server generates evaluation results based on the calculated degree of similarity. These evaluation results are output as numerical values ​​and comments indicating the reliability of the news articles. The server provides these evaluation results to users via a web interface or API.

[0552] (Application Example 1)

[0553] Next, we will describe Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0554] In modern society, many news articles are distributed via the internet, but some of them contain unreliable or misleading information. This creates a risk that readers may believe false information. Therefore, it is necessary to quickly and accurately evaluate the reliability of news articles and provide this information to readers.

[0555] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0556] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, and means for comparing with expert comments. This makes it possible to quickly and accurately evaluate the reliability of news articles and provide readers with reliable information.

[0557] A "news article" is a piece of writing or news report created to convey information, and is primarily distributed through the internet or print media.

[0558] "Reliability" refers to the degree to which information and data are judged to be accurate and free from errors.

[0559] An "information network" is a general term for communication systems, including the internet, that are used to exchange data and information with one another.

[0560] A "primary source" refers to a document or literature that provides the origin of information or the first recorded data.

[0561] "Means of verifying consistency" refers to methods and techniques for comparing the content of a news article with information from primary sources to determine whether they are consistent.

[0562] "Means of fact-checking" refers to methods and techniques for verifying whether the information contained in a news article is accurate.

[0563] "Expert comments" refer to opinions and explanations provided by experts who possess knowledge and experience in a particular field.

[0564] "Methods for cross-referencing" refers to methods and techniques for comparing the content of news articles with expert comments to verify their consistency.

[0565] An "information terminal" refers to electronic devices such as smartphones and tablets that are used to display and operate information.

[0566] A "generative AI model" refers to an algorithm or system that uses artificial intelligence technology to analyze data and is trained to perform a specific task.

[0567] A "confidence score" is a numerical indicator that quantifies the reliability of a news article and is used to evaluate the accuracy of the information.

[0568] The system for implementing this invention uses an information terminal and a server to evaluate the reliability of news articles. The information terminal is an electronic device such as a smartphone or tablet, which is used by the user when viewing news articles. The server performs the central processing for evaluating the reliability of news articles.

[0569] The server receives the text of news articles and uses bibliographic information on the information network to verify their match with primary sources. This involves using APIs to access publicly available databases on the internet (e.g., Google Scholar, PubMed). Furthermore, the server uses a generative AI model to analyze the content of the news articles and calculate a confidence score. This generative AI model is built using machine learning frameworks such as TensorFlow.

[0570] Once a confidence score is calculated, the server sends the result to the information terminal and displays it to the user. If the confidence score is low, the information terminal displays a warning to the user to draw their attention. This allows the user to immediately judge the reliability of the news article.

[0571] For example, if a user views a news article about the effectiveness of COVID-19 vaccines on their smartphone, the server evaluates the article's reliability and sends a reliability score of "85 / 100" to the user's device. If the reliability score is low (e.g., 40 / 100), the device displays a warning message.

[0572] Examples of prompts for a generative AI model include the following:

[0573] "Analyze the content of the news article and compare it against the following databases: Google Scholar, PubMed. Calculate the article's confidence score and output the result."

[0574] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0575] Step 1:

[0576] A user views a news article on an information terminal. The information terminal retrieves the text of the news article the user is viewing and sends that data to a server. The input is the text of the news article, and the output is the transmission of data to the server.

[0577] Step 2:

[0578] The server analyzes the text of the received news articles. Specifically, it uses natural language processing (NLP) techniques to tokenize the article's content and extract important keywords and phrases. The input is the text of the news article, and the output is the extracted keywords and phrases.

[0579] Step 3:

[0580] The server uses the extracted keywords and phrases to refer to publicly available databases on the information network (e.g., Google Scholar, PubMed) and verify their match with primary sources. It uses an API to query the databases and retrieve relevant bibliographic information. The input is the extracted keywords and phrases, and the output is the relevant bibliographic information.

[0581] Step 4:

[0582] The server uses a generative AI model to calculate the confidence score of news articles. The generative AI model compares the content of the news articles with the acquired bibliographic information and evaluates the degree of agreement. The input is the content of the news articles and the bibliographic information, and the output is the confidence score.

[0583] Step 5:

[0584] The server sends the calculated confidence score to the information terminal. The information terminal displays the received confidence score to the user. If the confidence score is low, the information terminal displays a warning message to alert the user. The input is the confidence score, and the output is the display to the user.

[0585] (Example 2)

[0586] Next, we will describe Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0587] In today's information society, the internet is flooded with a vast amount of information, some of which is unreliable. Therefore, both information recipients and providers are required to quickly and accurately assess the reliability of information. However, traditional methods for verifying information reliability are time-consuming and labor-intensive.

[0588] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0589] In this invention, the server includes means for verifying the consistency of the information text with a primary source using information from a communication network, means for performing fact-checking, means for analyzing the information using an information processing model and generating relevant terms, means for searching external sources to verify the information, and means for analyzing the search results and determining the reliability of the information. This makes it possible to quickly and accurately evaluate the reliability of the information.

[0590] A "system for evaluating the reliability of information" is a device or method for determining the accuracy and reliability of information by using information from materials on a communication network against the text of the information and confirming its consistency with the primary source.

[0591] "Information on communication networks" refers to information sources such as documents, articles, and databases that exist on the internet and other digital networks, and involves using these sources to verify and confirm information.

[0592] A "primary source" refers to a reliable source that provides the origin of information or original data, and serves as a standard for verifying the accuracy of information.

[0593] An "information processing model" refers to an algorithm or program used to analyze information and generate related terms, making the content of the information easier to understand.

[0594] "External information sources" refer to information providers or databases that exist outside the system and are used to verify and confirm information.

[0595] "Means of analyzing search results" refers to methods and devices for evaluating information obtained from external sources and determining the reliability of that information.

[0596] This invention is a system for evaluating the reliability of information, with a server at its core. The server receives information input from the user and performs a series of processes to evaluate the reliability of that information.

[0597] First, the server analyzes the input information using a natural language processing library. Specifically, it uses natural language processing software such as "spaCy" to tokenize the information and extract important elements such as nouns and verbs. This analysis allows the server to identify the subject of the information and related words.

[0598] Next, the server uses a generative AI model to generate relevant keywords from the extracted information. Based on these generated keywords, the server searches external information sources such as the "Google News API" and the "Bing Search API" to collect data to verify the reliability of the information.

[0599] The server analyzes search results obtained from external sources and determines the reliability of the information. Specifically, it compares search results and evaluates whether the information can be verified by other reliable sources. This evaluation result is provided to the user to help them quickly and accurately determine the reliability of the information.

[0600] For example, if a user enters "an article about the announcement of a new technology," the server analyzes the article and extracts keywords such as "technology" and "announcement." Next, it uses these keywords to search external sources and retrieve relevant news articles and official announcements. Finally, the server analyzes the search results and reports to the user whether the content of the article could be verified by other reliable sources.

[0601] An example of a prompt message is, "Please check if the information in this article can be verified by other reliable sources." This prompt allows users to efficiently evaluate the reliability of information.

[0602] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0603] Step 1:

[0604] The user inputs information they want to evaluate the reliability of from their terminal into the system. The server receives this input information and tokenizes the text using the natural language processing library "spaCy". Specifically, it breaks down the information into words and phrases and extracts important elements such as nouns and verbs. This process identifies the subject and related words of the information. The input is information from the user, and the output is the extracted keywords.

[0605] Step 2:

[0606] The server uses a generative AI model to generate relevant keywords based on the keywords extracted in Step 1. These generated keywords are appropriately selected based on the content of the information. Specifically, the generative AI model understands the context of the information and adds highly relevant phrases. This process forms a foundation for a deeper understanding of the information's content. The input is the extracted keywords, and the output is the generated relevant keywords.

[0607] Step 3:

[0608] The server uses the keywords generated in step 2 to search external information sources. Specifically, it uses APIs such as "Google News API" and "Bing Search API" to retrieve relevant news articles and official announcements. This search collects data to verify the reliability of the information. The input is the generated related keywords, and the output is the search results obtained from the external information sources.

[0609] Step 4:

[0610] The server analyzes the search results obtained in step 3 and determines the reliability of the information. Specifically, it compares the search results and evaluates whether the information can be verified by other reliable sources. This evaluation determines the accuracy of the information and reports it to the user. The input is the search results from external sources, and the output is the evaluation result regarding the reliability of the information.

[0611] Step 5:

[0612] The server provides the user with the evaluation results obtained in step 4. Specifically, it notifies the user of its judgment on whether the information is reliable and, if necessary, provides detailed information sources. This notification allows the user to quickly and accurately determine the reliability of the information. The input is the evaluation result regarding the reliability of the information, and the output is the notification to the user.

[0613] (Application Example 2)

[0614] Next, we will describe application example 2 of form example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 as the "terminal".

[0615] In modern society, the internet contains a vast amount of news articles, some of which contain unreliable or misleading information. Making decisions based on such information can have significant consequences for individuals and society as a whole. Therefore, there is a need to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0616] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0617] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information resources on an information network, means for fact-checking, means for comparing with expert opinions, means for notifying the user of the evaluation results, means for analyzing the article content using a generative AI model, and means for generating prompt sentences and comparing them with the information sources. This makes it possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0618] A "news article" is a collection of information provided on the internet or in print, and includes reports on specific events or incidents.

[0619] "Reliability" is an indicator that shows that information is accurate and free from errors.

[0620] An "information network" is a system for acquiring, transmitting, and sharing information through the internet and other digital communication methods.

[0621] "Information resources" refer to databases, official announcements, and other reliable sources used to assess the credibility of news articles.

[0622] A "primary source" refers to an institution or organization that provides official announcements or data directly related to the events being reported.

[0623] "Fact-checking" is the process of verifying whether reported information is accurate by comparing it with other reliable sources.

[0624] "Expert opinion" refers to evaluations or comments provided by individuals or organizations with knowledge or experience in a particular field.

[0625] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to generate and analyze text.

[0626] A "prompt statement" is an instruction given to a generative AI model, containing instructions for performing a specific task.

[0627] "User" refers to an individual or organization that receives the results of a reliability assessment of a news article.

[0628] The system for implementing this invention uses a server and a user terminal to evaluate the reliability of news articles. The server analyzes the text of the news article and uses information resources on the information network to verify its match with the primary source. Specifically, the server analyzes the article content using a generative AI model, generates prompt sentences, and compares them with the source. This process makes it possible to evaluate the reliability of the article.

[0629] The server notifies the user terminal of the evaluation results. The user terminal receives the evaluation results and displays them to the user. This allows the user to quickly verify the reliability of the news article.

[0630] As a concrete example, consider a scenario where a user is reading a news article stating that "the number of people infected with the new virus is rapidly increasing." The server extracts information about the number of infected people from the article and compares it with official announcements from the Ministry of Health, Labour and Welfare and reports from other reliable media outlets. Based on the comparison results, the server evaluates the reliability of the article as "high," "medium," or "low" and notifies the user's terminal.

[0631] Examples of prompts to input into a generative AI model include the following:

[0632] "Please compare the content of the following news article with reliable sources and assess its reliability. Article content: Article text"

[0633] In this way, by using a server and user terminals, it is possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0634] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0635] Step 1:

[0636] A user views a news article. The user's device sends the text data of the news article being viewed to the server. The input is the text data of the news article, and the output is the transmission to the server.

[0637] Step 2:

[0638] The server analyzes the text data of the received news articles. Using a generative AI model, it extracts important facts and information from the articles. The input is the text data of the news articles, and the output is the extracted facts and information.

[0639] Step 3:

[0640] The server generates a prompt message based on the extracted facts and information. The generated prompt message includes instructions for cross-referencing with reliable sources. The input is the extracted facts and information, and the output is the generated prompt message.

[0641] Step 4:

[0642] The server uses the generated prompt message to search for information resources on the information network and verify a match with the primary source. The input is the generated prompt message, and the output is the matching result.

[0643] Step 5:

[0644] The server evaluates the reliability of news articles based on the matching results. The evaluation is categorized as "high," "medium," or "low." The input is the matching results, and the output is the reliability evaluation result.

[0645] Step 6:

[0646] The server notifies the user terminal of the evaluation results. The user terminal displays the received evaluation results to the user. The input is the reliability evaluation result, and the output is the notification to the user.

[0647] In this way, it is possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[0648] (Example 3)

[0649] Next, we will describe Embodiment 3 of Embodiment Example 3. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0650] When evaluating the reliability of an informational article, it is necessary to efficiently check how well the article's content aligns with expert opinions and primary sources. However, traditional methods have the drawback of requiring recipients of information to spend a lot of time and effort judging the reliability of an article.

[0651] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 3 is realized by the following means.

[0652] In this invention, the server includes means for verifying the consistency of the information article text with primary sources using information from materials on an information network, means for verifying facts, and means for comparing with expert opinions. This makes it possible to quickly and accurately evaluate the reliability of the information article.

[0653] An "information article" is a piece of text, such as news or reports, provided on an information network.

[0654] "Reliability" refers to the characteristic that indicates information is accurate and free from errors and biases.

[0655] An "information network" is a system for acquiring, transmitting, and sharing information via the internet or other means.

[0656] A "primary source" is the origin of information and refers to documents or evidence that provide direct data or facts.

[0657] "Expert opinion" refers to views or comments provided by individuals with advanced knowledge and experience in a particular field.

[0658] "Natural language processing technology" refers to technologies that enable computers to understand, interpret, and generate human language.

[0659] A "data set" is a collection of data gathered for a specific purpose.

[0660] "Evaluation results" refer to information obtained as a result of evaluating the reliability and accuracy of the content of an informational article.

[0661] An "information terminal" is a device used for inputting, displaying, and processing information, and includes computers, smartphones, and other similar devices.

[0662] The following system configuration is used as an embodiment for carrying out this invention.

[0663] The server runs a program to evaluate the reliability of informational articles. This program analyzes the content of the informational articles using natural language processing techniques. Specifically, it uses a generative AI model to tokenize the content of the articles and extract important technical terms and claims. The server then compares the analyzed content of the informational articles with a data set containing expert opinions. This data set includes opinions from experts in various fields and is obtained from materials on the information network.

[0664] The terminal is responsible for sending informational article data entered by the user to the server. Users request analysis from the system by entering the URL or text of the article they want to analyze into the terminal. For example, if a user enters the prompt "I want to check opinions on the latest medical research," the terminal will send that information to the server.

[0665] The server generates and returns to the information terminal the results of its assessment of the reliability of the informational article. The assessment results include information about the article's reliability and its degree of agreement with expert opinions. Users can review these results through their terminal and determine the reliability of the informational article.

[0666] This system allows users to quickly and accurately verify how well the content of an informational article matches expert opinions and primary sources. The specific processing flow in Example 3 will be explained using Figure 15.

[0667] Step 1:

[0668] The user enters the URL or text of the information article they want to analyze into the terminal. The entered data is saved on the terminal as a prompt message. For example, the user might enter the prompt message, "I want to see opinions on the latest medical research."

[0669] Step 2:

[0670] The terminal sends the prompt text entered by the user to the server. The data sent includes the URL and text of the informational article, which serves as input data for the server to perform analysis.

[0671] Step 3:

[0672] The server analyzes the content of the informational article using natural language processing techniques based on the received prompt message. Specifically, it uses a generative AI model to tokenize the article's content and extract important technical terms and arguments. This process outputs the article's content as structured data.

[0673] Step 4:

[0674] The server compares the content of the analyzed informational article with a data set containing expert opinions. This data set includes opinions from experts in various fields and is obtained from materials on the information network. The degree of agreement between the article's claims and the expert opinions is then evaluated.

[0675] Step 5:

[0676] The server generates evaluation results and sends them back to the information terminal. These results include information about the article's reliability and its degree of agreement with expert opinions. This result becomes the output data from the server to the terminal.

[0677] Step 6:

[0678] The terminal displays the evaluation results received from the server to the user. Through the terminal, the user can check the reliability of the article and determine how well the content of the informational article matches expert opinions and primary sources.

[0679] (Application Example 3)

[0680] Next, we will describe application example 3 of form example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 as a "terminal".

[0681] While the internet contains a vast amount of news articles and information, it also includes misinformation and unreliable sources. This can lead readers and information recipients to make incorrect judgments, highlighting the need to efficiently evaluate the reliability of news articles and provide accurate information.

[0682] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means.

[0683] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, means for comparing with expert comments, means for analyzing the content of the article using natural language processing technology, and means for comparing with expert opinions and reliable sources using a generative AI model. This makes it possible to evaluate the reliability of news articles with high accuracy and provide accurate information to readers and information recipients.

[0684] A "news article" is a collection of information provided on the internet or in print media, and includes reports about specific events or incidents.

[0685] "Reliability" is an indicator that shows that information is accurate and free from errors.

[0686] A "primary source" refers to the original source or data from which information was first disseminated.

[0687] "Fact-checking" is the process of verifying whether the information provided is accurate.

[0688] "Expert comments" refer to opinions or views expressed by individuals with specialized knowledge in a particular field.

[0689] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[0690] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new information or data.

[0691] "Comparison" is the act of comparing two or more pieces of information or data and evaluating their similarities and differences.

[0692] The system for implementing this invention uses a server and a user terminal to evaluate the reliability of news articles. The server compares the text of the news article with literature information on the internet and verifies its consistency with primary sources. Furthermore, it performs fact-checking and cross-references it with expert comments. This makes it possible to evaluate the reliability of news articles with high accuracy.

[0693] The server analyzes the content of articles using natural language processing techniques. Specifically, it uses a natural language processing library (e.g., spaCy) to extract keywords and arguments from the articles. Next, it uses a generative AI model (e.g., OpenAI's GPT-3) to compare the extracted information with expert opinions and reliable sources. This comparison is used to evaluate the reliability of the articles.

[0694] The user's terminal receives evaluation results provided by the server and displays them to the user. This allows the user to quickly check the reliability of an article and reduces the risk of being misled by misinformation.

[0695] As a concrete example, if a user is reading an article about the effectiveness of a new vaccine, the server analyzes the claims in the article and inputs a prompt message into an AI model asking, "How do medical experts evaluate the effectiveness of this vaccine?" The system then compares the expert opinions returned by the model with the content of the article and evaluates their reliability.

[0696] The flow of the specific processing in Application Example 3 will be explained using Figure 16.

[0697] Step 1:

[0698] A user views a news article. The user's device sends the text data of the news article being viewed to the server. The input is the text data of the news article, and the output is the transmission of data to the server.

[0699] Step 2:

[0700] The server analyzes the text data of received news articles using natural language processing techniques. Specifically, it uses a natural language processing library (e.g., spaCy) to extract keywords and main arguments from the articles. The input is the text data of the news articles, and the output is the extracted keywords and main arguments.

[0701] Step 3:

[0702] The server inputs prompt sentences into a generative AI model (e.g., OpenAI's GPT-3) based on the extracted keywords and claims. These prompt sentences are in the form of seeking expert opinions related to the article's content. The input consists of extracted keywords and claims, and the output is the generated prompt sentences.

[0703] Step 4:

[0704] The generative AI model generates expert opinions and information from reliable sources based on the input prompt sentence. The input is the prompt sentence, and the output is the generated expert opinion or information.

[0705] Step 5:

[0706] The server compares expert opinions and information obtained from the generated AI model with the content of the news article. The comparison evaluates the reliability of the article. The input is the content of the news article and the generated expert opinions, and the output is the reliability evaluation result.

[0707] Step 6:

[0708] The server sends the reliability evaluation results to the user terminal. The user terminal displays the received evaluation results to the user. The input is the reliability evaluation results, and the output is the display to the user.

[0709] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0710] "Example of form 1"

[0711] One embodiment of the present invention provides a news article reliability evaluation system incorporating an emotion engine. This system includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes an emotion engine that recognizes the user's emotions. Specifically, it estimates emotions from the user's facial expressions and tone of voice while reading the article, as well as keystrokes when typing text, and reflects the results in the reliability evaluation. For example, if the user shows anger or distrust, the system lowers the reliability of the article. Conversely, if the user shows satisfaction or trust, the system increases the reliability of the article.

[0712] "Example of form 2"

[0713] Another embodiment of the present invention provides a news article reliability evaluation system incorporating an emotion engine. This system includes means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes means for providing the evaluation results to the article's readers and an emotion engine that recognizes the user's emotions. Specifically, the emotion engine recognizes the emotions the user feels when reading the article and adjusts the reliability evaluation based on those emotions. For example, if the user feels joy or excitement while reading the article, the reliability of that article is evaluated as high.

[0714] "Example of form 3"

[0715] As a further embodiment of the present invention, a news article reliability evaluation system incorporating an emotion engine is provided. This system includes means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes means for providing the evaluation results to the article writer and an emotion engine that recognizes the user's emotions. Specifically, the emotion engine recognizes the emotions the article writer felt when writing the article and adjusts the reliability evaluation based on those emotions. For example, if the writer felt anxiety or tension when writing the article, the reliability of that article will be evaluated as low.

[0716] The following describes the processing flow for each example of the form.

[0717] "Example of form 1"

[0718] Step 1: The user selects a news article.

[0719] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[0720] Step 3: The system verifies the facts.

[0721] Step 4: The system compares the information with expert comments.

[0722] Step 5: The emotion engine recognizes the user's emotions.

[0723] Step 6: The system incorporates the results of the emotion engine into the reliability assessment.

[0724] "Example of form 2"

[0725] Step 1: The user selects a news article.

[0726] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[0727] Step 3: The system verifies the facts.

[0728] Step 4: The system compares the information with expert comments.

[0729] Step 5: The emotion engine recognizes the user's emotions.

[0730] Step 6: The system incorporates the results of the emotion engine into the reliability assessment and provides the assessment results to the article's readers.

[0731] "Example of form 3"

[0732] Step 1: The writer creates the news article.

[0733] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[0734] Step 3: The system verifies the facts.

[0735] Step 4: The system compares the information with expert comments.

[0736] Step 5: The emotion engine recognizes the emotions of the article's writer.

[0737] Step 6: The system incorporates the results of the emotion engine into the reliability assessment and provides the assessment results to the article writer.

[0738] (Example 1)

[0739] Next, we will describe Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0740] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, with the vast amount of information available on the internet, verifying whether an article's content aligns with primary sources is not easy. Furthermore, while considering readers' sentiments and expert opinions is necessary when evaluating article reliability, there is a lack of efficient means to do so.

[0741] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0742] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, and means for recognizing user sentiment and reflecting the results in reliability evaluation. This makes it possible to evaluate the reliability of news articles from multiple perspectives and provide accurate evaluation results.

[0743] A "news article" refers to written text or news content created to convey information, and is particularly widely distributed through the internet and print media.

[0744] "Reliability" is an indicator that shows that information or data is accurate and error-free, and it represents the degree to which the recipient of the information can trust its content.

[0745] An "information network" refers to a communication infrastructure for sending and receiving digital data, including the internet, and is a system that allows access to various information sources.

[0746] A "primary source" refers to the place where specific information or data first originated, or the entity that directly provides that information, and forms the basis of reliable information.

[0747] "Fact-checking" is the process of verifying whether the information or data provided matches the actual facts, and is carried out to guarantee the accuracy of the information.

[0748] "Expert opinions" refer to views and analyses provided by individuals or organizations with advanced knowledge and experience in a particular field, and serve as a reference when evaluating the reliability of information.

[0749] "User emotions" refer to the psychological reactions and feelings that users exhibit when reading news articles, and are inferred from their facial expressions, tone of voice, input behavior, etc.

[0750] "Reliability assessment" is an evaluation process used to determine how accurate and reliable the content of a news article is, and it is carried out by considering various factors.

[0751] This invention is a system for evaluating the reliability of news articles, in which the server, terminal, and user elements work together.

[0752] When the server receives a news article, it analyzes the article text using natural language processing techniques. Specifically, the server tokenizes the text and extracts important keywords and numerical data. Common natural language processing libraries can be used for this analysis. Next, the server uses databases on the information network to compare the extracted data with primary sources. Databases such as Google Scholar and PubMed are used to search for information that matches the article's content and calculate the degree of match.

[0753] The device senses the user's facial expressions, tone of voice, and keystrokes when they are typing text while reading news articles. This uses hardware such as a camera, microphone, and keyboard. The device collects this data in real time and estimates the user's emotions. The emotion engine analyzes this emotion data and sends the emotions the user is expressing to the server.

[0754] The server adjusts the reliability rating of news articles based on sentiment data received from the device. If the user expresses anger or distrust, the server will rate the article's reliability lower. Conversely, if the user expresses satisfaction or trust, the server will rate the article's reliability higher.

[0755] As a concrete example, consider a scenario where a user is reading a news article about the effectiveness of a vaccine for the novel coronavirus. The server searches the information network for the data "95% vaccine effectiveness" found in the article and checks if it matches the primary source. Simultaneously, the terminal senses the user's facial expressions and tone of voice, and if the user shows distrust, the server lowers the reliability of the article.

[0756] An example of a prompt to input into a generative AI model is: "Explain how to evaluate the reliability of a news article by comparing the data in the article with primary sources on the information network and adjusting the reliability by considering user sentiment data." This prompt allows the generative AI model to generate a detailed explanation of the system's process.

[0757] The flow of the specific processing in Example 1 will be explained using Figure 17.

[0758] Step 1:

[0759] The server receives news articles from users. The input is the text of the news articles. The server analyzes the article text using natural language processing techniques and extracts important keywords and numerical data. Specifically, it tokenizes the text and identifies parts of speech such as nouns and verbs. The output is a list of the extracted keywords and data.

[0760] Step 2:

[0761] The server searches databases on the information network based on the data extracted in Step 1. Keywords and a list of data are used as input. The server uses databases such as Google Scholar and PubMed to find information that matches the primary sources. Specifically, it sends queries to the databases via APIs to retrieve relevant literature and data. A score indicating the degree of matching is generated as output.

[0762] Step 3:

[0763] The device collects user emotion data when the user reads news articles. Input includes the user's facial expressions, voice tone, and keystrokes during text input. The device uses its camera, microphone, and keyboard to sense and collect this data in real time. Specifically, it uses facial recognition and voice analysis technologies to estimate the user's emotions. The estimated emotion data is then generated as output.

[0764] Step 4:

[0765] The server integrates the similarity score obtained in step 2 with the sentiment data obtained in step 3 to evaluate the reliability of the news article. The similarity score and sentiment data are used as input. The server uses a sentiment engine to calculate the impact of the user's emotions on the reliability evaluation. Specifically, if anger or distrust is indicated, the reliability is rated low, and if satisfaction or trust is indicated, the reliability is rated high. The final reliability evaluation score is generated as output.

[0766] Step 5:

[0767] The server presents the final reliability rating score to the user. The reliability rating score is used as input. The server displays the evaluation results visually, making them easy for the user to understand. Specifically, the reliability score is shown numerically and graphically, providing the user with information to judge the reliability of the article. A visualized evaluation result is generated as output.

[0768] (Application Example 1)

[0769] Next, we will describe Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0770] In modern society, where the reliability of information is paramount, there is a need to quickly and accurately evaluate the reliability of news articles. However, conventional methods are time-consuming in verifying the accuracy of primary sources and fact-checking, and they do not take into account the emotions of users in their reliability assessments. As a result, it is difficult to provide information that is truly reliable to recipients.

[0771] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0772] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, and means for recognizing user sentiment and reflecting the results in the reliability evaluation. This enables rapid and accurate evaluation of the reliability of news articles and personalized reliability evaluations that take user sentiment into consideration.

[0773] A "news article" is a piece of writing intended to convey information, and in particular, it includes content related to current events and social issues.

[0774] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[0775] An "information network" refers to a communication system for sending and receiving digital data, including the internet.

[0776] A "primary source" refers to the source of information or documents and data that provide direct evidence.

[0777] "Fact-checking" refers to the process of verifying whether information is accurate.

[0778] "Expert opinion" refers to the views of individuals who possess advanced knowledge and experience in a particular field.

[0779] "User" refers to an individual or group that uses the system or service.

[0780] "Recognizing emotions" refers to the process of estimating a user's emotional state from their facial expressions, tone of voice, and behavior.

[0781] "Reliability assessment" refers to the process of demonstrating the accuracy and reliability of information using numerical values ​​and indicators.

[0782] "Personalization" refers to adjusting services and information according to the individual characteristics and preferences of each user.

[0783] The system for implementing this invention involves a server and a terminal working together to evaluate the reliability of news articles. The server receives the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. This verifies the facts of the article and compares them with expert opinions.

[0784] When a user views a news article, the device uses its camera and microphone to capture the user's facial expressions and tone of voice. This allows the device to recognize emotions and send the results to a server. The server then adjusts its reliability assessment based on the received emotion data and calculates a final reliability score.

[0785] This system will be implemented as an application installed on devices such as smartphones and smart glasses. Specifically, it will perform natural language processing and sentiment analysis using software such as Python and TensorFlow. spaCy will be used as the natural language processing library, and NLTK will be used as the sentiment analysis library.

[0786] For example, when a user is reading a news article through smart glasses, the article's reliability score is displayed in their field of vision. If the user frowns, the emotion engine detects the distrust and adjusts the reliability score.

[0787] An example of a prompt to input into a generative AI model is: "Please rate the reliability of this news article. The article content is as follows: 'Article Content'. The user's sentiment indicates distrust."

[0788] The flow of a specific process in Application Example 1 will be explained using Figure 18.

[0789] Step 1:

[0790] The device uses its camera and microphone to capture the user's facial expressions and voice tone when they view news articles. This data is used as input for sentiment analysis.

[0791] Step 2:

[0792] The device inputs the acquired facial and voice data into an emotion analysis library (such as NLTK) to estimate the user's emotions. The result of the emotion analysis is output, indicating the emotion the user is expressing (e.g., distrust, satisfaction).

[0793] Step 3:

[0794] The terminal sends the text of the news article to the server. The server analyzes the received article using a natural language processing library (such as spaCy) and extracts facts and data from the article.

[0795] Step 4:

[0796] The server compares the extracted facts and data with bibliographic information on the information network and calculates the degree of match with the primary source. This degree of match is output as the basis for reliability evaluation.

[0797] Step 5:

[0798] The server compares the article content with an expert opinion database and evaluates the degree of agreement between the expert's view and the article. This evaluation result is also taken into account in the reliability assessment.

[0799] Step 6:

[0800] The server receives user sentiment data and incorporates it into the reliability assessment. Specifically, if a user expresses distrust, the reliability score is adjusted.

[0801] Step 7:

[0802] The server calculates the final reliability score and sends it to the terminal. The terminal then displays this score visually to the user.

[0803] Step 8:

[0804] Users can check the displayed reliability score to determine the reliability of a news article. This allows users to read articles while considering the accuracy of the information.

[0805] (Example 2)

[0806] Next, we will describe Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0807] When evaluating the reliability of news articles, there is a challenge in objectively judging the accuracy and reliability of the information. Furthermore, since the influence of user emotions on the reliability evaluation of an article is not taken into account, the evaluation results may be influenced by the user's subjectivity. In addition, there is a lack of means to visually present the evaluation results, making it difficult for users to intuitively understand the results.

[0808] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0809] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, means for recognizing user sentiment and adjusting reliability evaluation based on that sentiment, and means for visually displaying the evaluation results. This enables objective and intuitive evaluation of the reliability of news articles, and allows for reliability evaluation that takes user sentiment into consideration.

[0810] A "news article" is written or digital content created to report events or information.

[0811] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[0812] An "information network" is a system for acquiring, transmitting, and sharing information through the internet and other digital communication methods.

[0813] A "primary source" refers to the direct origin or source of an event or piece of information.

[0814] "Fact-checking" is the process of verifying whether information is accurate using other reliable sources.

[0815] "Expert opinion" refers to the views of an individual or organization that possesses advanced knowledge and experience in a particular field.

[0816] "User sentiment" refers to the psychological reactions that users exhibit when reading news articles.

[0817] "Visual display methods" refer to methods of presenting information to users in an easy-to-understand manner using graphs, diagrams, numerical data, etc.

[0818] As an embodiment of this invention, a system for evaluating the reliability of news articles is constructed. The server receives news articles and analyzes their content using natural language processing technology. Specifically, it uses a text analysis library (e.g., spaCy or NLTK) to extract important facts and keywords from the articles.

[0819] The server searches for reliable sources of information through its information network based on extracted facts and keywords. Using APIs, it accesses official police announcements and other media databases to verify the facts presented in the articles. It also sends queries to external databases using HTTP requests.

[0820] The device recognizes the user's emotions as they read the article. It uses the camera and microphone to capture the user's facial expressions and voice, which are then analyzed by an emotion engine. The emotion engine uses machine learning models (e.g., TensorFlow or PyTorch) to classify the user's emotions into categories such as "joy," "anxiety," and "excitement."

[0821] The server integrates fact-checking results with user sentiment assessment results to evaluate the reliability of news articles. If a user expresses positive emotions after reading an article, the reliability score increases. Conversely, if a user expresses negative emotions, more emphasis is placed on cross-referencing with other sources.

[0822] Users receive the reliability evaluation results through their device. The device visually displays the evaluation results, providing the article's reliability score and comments. The evaluation results are presented in graphs and numerical values, in a way that is easy for users to understand.

[0823] As a concrete example, when a user enters a news article about the spread of a new virus, the server analyzes the article's content and verifies relevant facts using external sources. The device captures the user's facial expressions with its camera and analyzes them with an emotion engine. Finally, the reliability evaluation results are displayed on the device, allowing the user to verify the article's reliability.

[0824] Examples of prompts for a generative AI model:

[0825] "Analyze news articles about the spread of the new virus and evaluate their reliability. Verify the facts in the articles with other reliable sources and adjust your evaluation considering user sentiment."

[0826] The flow of the specific processing in Example 2 will be explained using Figure 19.

[0827] Step 1:

[0828] The server receives news articles from users as input. The server analyzes the article content using natural language processing techniques, extracting important facts and keywords. This process utilizes text analysis libraries (e.g., spaCy or NLTK) to analyze the article's grammatical structure and identify keywords. The server then generates a list of extracted keywords and facts as output.

[0829] Step 2:

[0830] The server uses the keywords and facts extracted in Step 1 as input to search for reliable sources of information through the information network. The server accesses external databases using APIs to verify whether the facts in the articles can be confirmed. Specifically, it sends queries to external databases using HTTP requests and retrieves the matching results as output.

[0831] Step 3:

[0832] The device recognizes the user's emotions as input while reading news articles. The device uses its camera and microphone to capture the user's facial expressions and voice, which are then analyzed by an emotion engine. The emotion engine uses machine learning models (e.g., TensorFlow or PyTorch) to classify the user's emotions into categories such as "joy," "anxiety," and "excitement." As output, it generates user emotion data.

[0833] Step 4:

[0834] The server integrates the matching results from Step 2 and the sentiment data from Step 3 as input to evaluate the reliability of the news article. If the user expresses positive sentiment after reading the article, the server increases the reliability score. Conversely, if the user expresses negative sentiment, it places more emphasis on the matching results with other sources. The server generates a reliability evaluation score as output.

[0835] Step 5:

[0836] The user receives the reliability evaluation results through their device. The device visually displays the evaluation results and provides the article's reliability score and comments. Specifically, it presents the evaluation results in graphs and numerical values ​​in a way that is easy for the user to understand. As output, it generates a visually displayed reliability evaluation result.

[0837] (Application Example 2)

[0838] Next, we will describe application example 2 of form example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 as the "terminal".

[0839] When evaluating the reliability of news articles, it is necessary to consider not only the content but also the reader's emotions. However, conventional systems do not perform reliability evaluations that take emotions into account, which presents a challenge in providing readers with reliable information.

[0840] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0841] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, and means for recognizing the user's emotions and adjusting the reliability evaluation based on those emotions. This makes it possible to evaluate the reliability of news articles while taking the reader's emotions into consideration and to provide highly reliable information.

[0842] A "news article" refers to written text or news reports created to convey information.

[0843] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[0844] An "information network" refers to a communication network used to transmit information, including the internet.

[0845] A "primary source" is a source that provides the origin of information or direct evidence.

[0846] "Fact-checking" is the process of verifying whether information is accurate.

[0847] "Expert comments" refer to opinions or explanations provided by individuals with specialized knowledge in a particular field.

[0848] "Users" refer to individuals or organizations that use a system or service.

[0849] "Emotion" refers to a person's psychological reaction or state, and includes feelings such as joy and sadness.

[0850] "Evaluation" refers to the act of judging value or performance based on specific criteria.

[0851] A "server" is a computer system that provides information and services over a network.

[0852] The system for carrying out this invention includes a program for evaluating the reliability of news articles. The server analyzes the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. This is done using natural language processing libraries (e.g., NLTK, spaCy). Furthermore, it uses external APIs (e.g., Google Fact Check Tools API) to verify the accuracy of the article for fact-checking.

[0853] The device acquires data through its camera and microphone to recognize the user's emotions and analyzes it using an emotion recognition library (e.g., Affectiva). This allows for adjusting the reliability evaluation based on the user's emotions. The evaluation results are displayed on the device's screen and provided to the user.

[0854] As a concrete example, when a user opens a news app on their smartphone and begins reading an article, the server evaluates the article's reliability and displays a reliability score on the screen. If the user smiles while reading the article, the device recognizes that emotion and adjusts the reliability score accordingly.

[0855] An example of a prompt to input into a generative AI model is: "Please evaluate the reliability of this news article. Compare the article's content with external sources and calculate a reliability score, taking into account the user's sentiment."

[0856] The flow of a specific process in Application Example 2 will be explained using Figure 20.

[0857] Step 1:

[0858] The server retrieves the text of the news article selected by the user. It receives the URL or ID of the news article as input and extracts the text data of the article. It generates the text data of the news article to be analyzed as output.

[0859] Step 2:

[0860] The server analyzes the text data of news articles using natural language processing libraries (e.g., NLTK, spaCy). It receives the text data of news articles as input and extracts keywords and important phrases from the articles. As output, it generates data that summarizes the content of the articles.

[0861] Step 3:

[0862] The server uses external APIs (e.g., Google Fact Check Tools API) to fact-check the content of news articles. It receives summary data as input and compares it with external sources. As output, it generates initial assessment data regarding the reliability of the article.

[0863] Step 4:

[0864] The device acquires data on the user's facial expressions and voice through its camera and microphone in order to recognize the user's emotions. It receives real-time video and audio data of the user as input and analyzes it using an emotion recognition library (e.g., Affectiva). As output, it generates data indicating the user's emotional state.

[0865] Step 5:

[0866] The server adjusts the reliability rating of news articles based on the user's emotional state data. It receives initial evaluation data and emotional state data as input and recalculates the reliability score. It generates the adjusted reliability score as output.

[0867] Step 6:

[0868] The device displays the adjusted reliability score to the user. It receives the adjusted reliability score as input and displays it visually on the user's screen. As output, it provides a reliability evaluation result that the user can visually confirm.

[0869] (Example 3)

[0870] Next, we will describe Embodiment 3 of Embodiment Example 3. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0871] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, verifying whether an article's content aligns with primary sources and expert opinions is not easy. Furthermore, it is necessary to consider the impact of the writer's emotions on the article's reliability. An effective system is needed to address these challenges.

[0872] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 3 is realized by the following means.

[0873] In this invention, the server includes means for analyzing information, means for confirming matching with information sources, and means for verifying facts. This makes it possible to evaluate the reliability of news articles from multiple perspectives and to quickly and accurately determine their reliability.

[0874] "Means of analyzing information" refers to a function that uses natural language processing technology to analyze the content of news articles and extract expert arguments and keywords.

[0875] "Means of verifying consistency with information sources" refers to the function of searching relevant literature and databases and comparing their contents to confirm whether the claims in a news article are consistent with primary sources.

[0876] "Means of fact-checking" refers to the function of verifying the accuracy of information contained in news articles by referring to relevant data and expert opinions.

[0877] "Means of comparing with expert opinions" refers to a function that compares and evaluates the claims made in news articles with the opinions and comments of experts in that field.

[0878] "Means of recognizing emotions" refers to the function of analyzing the emotions of news article writers and evaluating the impact of those emotions on the article's credibility.

[0879] "Means for adjusting evaluations" refers to a function that adjusts the reliability evaluation of news articles based on collected information and sentiment analysis results, and generates a final evaluation result.

[0880] To implement this invention, it is necessary to build a system for evaluating the reliability of news articles. The user inputs the news article they wish to evaluate into the system. The server analyzes the content of the article using natural language processing techniques and extracts expert claims and keywords. This analysis uses the Python natural language processing library.

[0881] Next, the server gathers relevant literature and expert opinions from the internet based on the extracted claims. Specifically, it uses academic paper database APIs to search for relevant literature and compare it with the claims in the article. Furthermore, the server uses database query techniques to verify that the claims in the article match those of primary sources.

[0882] Furthermore, the server uses an emotion engine to analyze the emotions of the article's writer. Specifically, it uses a Python emotion analysis library to determine whether the writer is feeling anxious or stressed. This emotion information influences the reliability evaluation.

[0883] Finally, the server integrates the collected information and sentiment analysis results to evaluate the reliability of the news article. The evaluation results are displayed visually to the user. For example, the reliability score may be shown in a graph, with a detailed explanation of which factors influenced the evaluation.

[0884] As a concrete example, when a user enters "an article about the effectiveness of a new virus vaccine," the server analyzes the article and extracts claims about vaccine effectiveness. Next, the server searches for relevant medical papers using an academic paper database and evaluates the accuracy of the claims. Simultaneously, an emotion engine analyzes the writer's emotions and incorporates this into the reliability assessment. Finally, the server generates a reliability score and provides it to the user.

[0885] An example of a prompt to the generative AI model might be: "Please verify whether the medical claims in this article are consistent with expert opinions. Also, please evaluate the impact of the writer's emotions on its credibility." The specific processing flow in Example 3 will be explained using Figure 21.

[0886] Step 1:

[0887] The user inputs a news article they want to evaluate into the system. The server receives the input article and analyzes its content using natural language processing techniques. Specifically, it uses a Python natural language processing library to extract expert claims and keywords from the article. As a result of this analysis, a list of the article's claims and keywords is output.

[0888] Step 2:

[0889] Based on the claims extracted in Step 1, the server collects relevant literature and expert opinions from the internet. Specifically, it uses an academic paper database API to search for relevant literature and compares it with the claims in the article. This process takes a list of claims as input and produces a list of relevant literature and expert comments as output.

[0890] Step 3:

[0891] The server verifies whether the claims in the article match those of the primary source. Specifically, it uses database query techniques to identify the primary source and check for content consistency. This step takes a list of claims as input and outputs results indicating whether or not there is a match.

[0892] Step 4:

[0893] The server uses an emotion engine to analyze the emotions of the article's author. Specifically, it uses a Python emotion analysis library to determine whether the author is feeling anxious or stressed. This process takes the article's text as input and outputs the results of the emotion analysis.

[0894] Step 5:

[0895] The server integrates the collected information and sentiment analysis results to evaluate the reliability of the news article. Specifically, it calculates a reliability score and generates an evaluation result. In this step, relevant literature, expert comments, agreement results, and sentiment analysis results are used as input, and the reliability score is obtained as output.

[0896] Step 6:

[0897] The server provides the user with the final evaluation results. Specifically, it visually displays the reliability score and provides a detailed explanation of which factors influenced the evaluation. In this step, the reliability score is used as input, and the evaluation results are displayed to the user as output.

[0898] (Application Example 3)

[0899] Next, we will describe application example 3 of form example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 as a "terminal".

[0900] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, verifying whether an article's content aligns with expert opinions and primary sources is not easy, nor is it feasible to consider the impact of the writer's emotions on its reliability. Therefore, there is a need for a system that comprehensively evaluates the reliability of news articles and provides this information to readers and writers.

[0901] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means.

[0902] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, means for analyzing the sentiment of the article's writer and adjusting the reliability evaluation, and means for displaying the evaluation results on a display device. This makes it possible to comprehensively evaluate the reliability of news articles and provide this information to readers and writers quickly and accurately.

[0903] A "news article" is a piece of writing created to convey information, and includes reporting on specific events or incidents.

[0904] "Reliability" is a concept that refers to the degree to which information is judged to be accurate and free from errors.

[0905] An "information network" is a system that uses digital communication technologies such as the internet to share information.

[0906] A "primary source" refers to the place or person from which specific information or data first originated; it is the original source of the information.

[0907] "Expert opinion" refers to views or comments from individuals who possess advanced knowledge and experience in a particular field.

[0908] "Sentiment analysis" is a technique that extracts emotions from written text or speech and evaluates the type and intensity of those emotions.

[0909] "Evaluation results" refer to the results of an evaluation based on specific criteria, and in this context, it means an evaluation of the reliability of a news article.

[0910] A "display device" is a device used to visually display information, and includes computer monitors and smartphone screens.

[0911] An "information terminal" is an electronic device used to process and display information, and includes smartphones and tablets.

[0912] The term "creator" refers to the person who produced a particular piece of writing or work.

[0913] The system for implementing this invention uses a server and an information terminal to evaluate the reliability of news articles. The server analyzes the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. Furthermore, it performs fact-checking and compares it with expert opinions. This makes it possible to comprehensively evaluate the reliability of the article.

[0914] The server uses a sentiment analysis API (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment of the article's author and adjust the reliability rating accordingly. The evaluation results are displayed on the information terminal's display device, allowing users to check them in real time.

[0915] As a concrete example, when a user opens a news app on their smartphone and selects a specific article, the article's reliability score is displayed on the screen. For instance, if the article reports on a "new medical discovery," the system checks whether the content aligns with the opinions of medical experts and calculates a reliability score.

[0916] An example of a prompt for a generative AI model is, "Verify whether the medical claims in this article are consistent with expert opinions and calculate a reliability score." This prompt prompts the server to perform the necessary data processing and reliability assessment.

[0917] The flow of the specific processing in Application Example 3 will be explained using Figure 22.

[0918] Step 1:

[0919] The user launches a news app on their information terminal and selects a specific article. The input is the news article selected by the user, and the output is the text data of that article. The terminal sends this text data to the server.

[0920] Step 2:

[0921] The server analyzes the text data of received news articles and uses bibliographic information on the information network to verify their match with primary sources. The input is the text data of news articles, and the output is the result of matching with primary sources. The server analyzes the text using a natural language processing library (e.g., spaCy) and verifies the match by calling an external fact-checking API.

[0922] Step 3:

[0923] The server verifies the content of news articles and compares them with expert opinions. The input is the text data of the news article, and the output is the results of the fact-checking and the agreement with expert opinions. The server refers to an expert database to check whether the content of the article is consistent with the opinions of experts.

[0924] Step 4:

[0925] The server uses a sentiment analysis API to analyze the writer's sentiment and adjust the reliability rating accordingly. The input is the text data of the news article, and the output is the result of the writer's sentiment analysis. The server calls a sentiment analysis API (e.g., IBM Watson Natural Language Understanding) to evaluate the writer's sentiment.

[0926] Step 5:

[0927] The server integrates these results to calculate the reliability score of the news article. The inputs are the results of matching primary sources, fact-checking, agreement with expert opinions, and sentiment analysis, and the output is the reliability score. The server integrates this data and runs an algorithm to calculate the reliability score.

[0928] Step 6:

[0929] The server sends the calculated reliability score to the information terminal, and the terminal displays the score to the user. The input is the reliability score, and the output is the reliability score displayed on the terminal's display device. The terminal displays the received score on the screen so that the user can verify it.

[0930] (Other examples)

[0931] Since this is the same as the specific processing described in the other embodiments of the first embodiment above, the explanation will be omitted.

[0932] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0933] The data generation model 58 is a form of so-called generative AI (Artificial Intelligence). One example of the data generation model 58 is ChatGPT (Internet Search).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0934] Other examples of generative AI include Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are examples.

[0935] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0936] [Third Embodiment]

[0937] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0938] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0939] The data processing device 12 includes a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a “computer” related to the technology of this disclosure.

[0940] Computer 22 includes a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. A database 24 and a communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0941] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0942] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0943] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0944] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0945] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0946] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0947] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0948] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0949] Next, the identification process performed by the identification processing unit 290 of the data processing device 12 will be described.

[0950] "Example of form 1"

[0951] One embodiment of the present invention is a system for evaluating the reliability of news articles. This system has a means for verifying the consistency between the text of a news article and its primary source using information from online sources. Specifically, it compares the facts and data described in the article with publicly available papers and databases on the internet and calculates the degree of agreement.

[0952] "Example of form 2"

[0953] Furthermore, the system of the present invention also includes means for fact-checking. This verifies whether the facts and information described in the article can be confirmed by other reliable sources. For example, it verifies whether the incident reported in the article has been reported in official police statements or by other media outlets.

[0954] "Example of form 3"

[0955] Furthermore, the system of the present invention includes means for cross-referencing with expert comments. This means that the specialized content and views described in the article are evaluated in comparison with the opinions and comments of experts in that field. For example, it checks whether the medical claims reported in the article are consistent with comments and papers from experts in the medical community.

[0956] The following describes the processing flow for each example of the form.

[0957] "Example of form 1"

[0958] Step 1: The system receives the news article text as input.

[0959] Step 2: Extract the facts and data contained in the article.

[0960] Step 3: Compare the extracted facts and data with publicly available papers and databases on the internet and calculate the degree of agreement.

[0961] "Example of form 2"

[0962] Step 1: The system receives the news article text as input.

[0963] Step 2: Extract the facts and information contained in the article.

[0964] Step 3: Verify whether the extracted facts and information can be confirmed by other reliable sources.

[0965] "Example of form 3"

[0966] Step 1: The system receives the news article text as input.

[0967] Step 2: Extract the specialized content and viewpoints mentioned in the article.

[0968] Step 3: Evaluate the extracted specialized content and views by comparing them with the opinions and comments of experts in that field.

[0969] (Example 1)

[0970] Next, we will describe Embodiment 1 of Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0971] There is a need to quickly and accurately assess the reliability of news articles, but traditional methods are time-consuming to cross-reference and evaluate information, making it difficult to obtain reliable results.

[0972] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0973] In this invention, the server includes means for verifying the match between the news article text and primary sources using information from literature on an information network, means for analyzing the article content using natural language processing technology, and means for searching publicly available sources on an information network and comparing facts and data within the article. This makes it possible to quickly and accurately evaluate the reliability of news articles.

[0974] A "news article" is a piece of writing intended to convey information, providing details about a specific event or occurrence.

[0975] "Reliability" is an indicator that shows that information is accurate and error-free, and it is a standard for evaluating the legitimacy and credibility of information.

[0976] An "information network" is a communication infrastructure for sending and receiving digital data, including the internet.

[0977] A "primary source" is the initial source of specific information or data, and is a source that provides direct evidence or data.

[0978] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and it involves analyzing text and extracting meaning.

[0979] "Public information sources" are databases and documents that provide information that is generally accessible to the public, and are collections of information that anyone can use.

[0980] A "text similarity calculation algorithm" is a computational method for numerically evaluating the similarity between different texts, and it measures the degree of textual agreement.

[0981] "Evaluation results" refer to the outcome of an evaluation conducted based on specific criteria, and indicate conclusions regarding reliability and accuracy.

[0982] To implement this invention, it is necessary to build a system for evaluating the reliability of news articles. The user sends the text of a news article to the server. The server analyzes the received news article using natural language processing technology. This analysis utilizes natural language processing libraries such as "spaCy" and "NLTK". The server extracts facts and data from the article and searches publicly available information sources on the information network. This search utilizes academic paper search APIs and database APIs.

[0983] The server compares the content of news articles with primary sources obtained through searches. For this comparison, it uses text similarity calculation algorithms such as "Cosine Similarity" and "Jaccard Index." The server uses these algorithms to calculate the degree of similarity and generate evaluation results. These evaluation results are provided to the user via a web interface or API.

[0984] As a concrete example, consider a case where a user wants to evaluate a news article about "the global economic growth rate in 2023." The server extracts the statement "the global economic growth rate in 2023 is 3.5%" from the article. Next, it uses an academic paper search API to search for related papers using the keyword "2023 global economic growth rate." The server compares the content of the papers obtained from the search results with the description in the article and calculates the degree of similarity using Cosine Similarity. Finally, the server evaluates the reliability as higher if the degree of similarity is high and returns the result to the user.

[0985] An example of a prompt to input into a generative AI model might be, "To evaluate the reliability of the news article, please compare the facts in the article with online sources."

[0986] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0987] Step 1:

[0988] The user sends the body of a news article to the server. The text data of the news article is provided as input. The server receives this text data and prepares for the next analysis step.

[0989] Step 2:

[0990] The server analyzes the text of received news articles using natural language processing techniques. Specifically, it uses the natural language processing library "spaCy" to extract facts and data from the articles. The input is the text data of the news article, and the output is a list of the extracted facts and data.

[0991] Step 3:

[0992] The server searches publicly available information sources on the information network based on the extracted facts and data. This search uses an academic paper search API. The input is a list of extracted facts and data, and the output is a list of relevant primary sources. Specifically, the server performs a search using the keyword "2023 global economic growth rate".

[0993] Step 4:

[0994] The server compares the content of news articles with primary sources obtained through searches. The "Cosine Similarity" algorithm, a text similarity calculation algorithm, is used for this comparison. The input consists of the text data of the news articles and a list of primary sources; the output is a numerical score indicating the degree of similarity. The server uses this to evaluate the reliability of the articles.

[0995] Step 5:

[0996] The server generates evaluation results based on the calculated degree of similarity. These evaluation results are output as numerical values ​​and comments indicating the reliability of the news articles. The server provides these evaluation results to users via a web interface or API.

[0997] (Application Example 1)

[0998] Next, we will describe Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0999] In modern society, many news articles are distributed via the internet, but some of them contain unreliable or misleading information. This creates a risk that readers may believe false information. Therefore, it is necessary to quickly and accurately evaluate the reliability of news articles and provide this information to readers.

[1000] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1001] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, and means for comparing with expert comments. This makes it possible to quickly and accurately evaluate the reliability of news articles and provide readers with reliable information.

[1002] A "news article" is a piece of writing or news report created to convey information, and is primarily distributed through the internet or print media.

[1003] "Reliability" refers to the degree to which information and data are judged to be accurate and free from errors.

[1004] An "information network" is a general term for communication systems, including the internet, that are used to exchange data and information with one another.

[1005] A "primary source" refers to a document or literature that provides the origin of information or the first recorded data.

[1006] "Means of verifying consistency" refers to methods and techniques for comparing the content of a news article with information from primary sources to determine whether they are consistent.

[1007] "Means of fact-checking" refers to methods and techniques for verifying whether the information contained in a news article is accurate.

[1008] "Expert comments" refer to opinions and explanations provided by experts who possess knowledge and experience in a particular field.

[1009] "Methods for cross-referencing" refers to methods and techniques for comparing the content of news articles with expert comments to verify their consistency.

[1010] An "information terminal" refers to electronic devices such as smartphones and tablets that are used to display and operate information.

[1011] A "generative AI model" refers to an algorithm or system that uses artificial intelligence technology to analyze data and is trained to perform a specific task.

[1012] A "confidence score" is a numerical indicator that quantifies the reliability of a news article and is used to evaluate the accuracy of the information.

[1013] The system for implementing this invention uses an information terminal and a server to evaluate the reliability of news articles. The information terminal is an electronic device such as a smartphone or tablet, which is used by the user when viewing news articles. The server performs the central processing for evaluating the reliability of news articles.

[1014] The server receives the text of news articles and uses bibliographic information on the information network to verify their match with primary sources. This involves using APIs to access publicly available databases on the internet (e.g., Google Scholar, PubMed). Furthermore, the server uses a generative AI model to analyze the content of the news articles and calculate a confidence score. This generative AI model is built using machine learning frameworks such as TensorFlow.

[1015] Once a confidence score is calculated, the server sends the result to the information terminal and displays it to the user. If the confidence score is low, the information terminal displays a warning to the user to draw their attention. This allows the user to immediately judge the reliability of the news article.

[1016] For example, if a user views a news article about the effectiveness of COVID-19 vaccines on their smartphone, the server evaluates the article's reliability and sends a reliability score of "85 / 100" to the user's device. If the reliability score is low (e.g., 40 / 100), the device displays a warning message.

[1017] Examples of prompts for a generative AI model include the following:

[1018] "Analyze the content of the news article and compare it against the following databases: Google Scholar, PubMed. Calculate the article's confidence score and output the result."

[1019] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1020] Step 1:

[1021] A user views a news article on an information terminal. The information terminal retrieves the text of the news article the user is viewing and sends that data to a server. The input is the text of the news article, and the output is the transmission of data to the server.

[1022] Step 2:

[1023] The server analyzes the text of the received news articles. Specifically, it uses natural language processing (NLP) techniques to tokenize the article's content and extract important keywords and phrases. The input is the text of the news article, and the output is the extracted keywords and phrases.

[1024] Step 3:

[1025] The server uses the extracted keywords and phrases to refer to publicly available databases on the information network (e.g., Google Scholar, PubMed) and verify their match with primary sources. It uses an API to query the databases and retrieve relevant bibliographic information. The input is the extracted keywords and phrases, and the output is the relevant bibliographic information.

[1026] Step 4:

[1027] The server uses a generative AI model to calculate the confidence score of news articles. The generative AI model compares the content of the news articles with the acquired bibliographic information and evaluates the degree of agreement. The input is the content of the news articles and the bibliographic information, and the output is the confidence score.

[1028] Step 5:

[1029] The server sends the calculated confidence score to the information terminal. The information terminal displays the received confidence score to the user. If the confidence score is low, the information terminal displays a warning message to alert the user. The input is the confidence score, and the output is the display to the user.

[1030] (Example 2)

[1031] Next, we will describe Example 2 of the morphological example. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1032] In today's information society, the internet is flooded with a vast amount of information, some of which is unreliable. Therefore, both information recipients and providers are required to quickly and accurately assess the reliability of information. However, traditional methods for verifying information reliability are time-consuming and labor-intensive.

[1033] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1034] In this invention, the server includes means for verifying the consistency of the information text with a primary source using information from a communication network, means for performing fact-checking, means for analyzing the information using an information processing model and generating relevant terms, means for searching external sources to verify the information, and means for analyzing the search results and determining the reliability of the information. This makes it possible to quickly and accurately evaluate the reliability of the information.

[1035] A "system for evaluating the reliability of information" is a device or method for determining the accuracy and reliability of information by using information from materials on a communication network against the text of the information and confirming its consistency with the primary source.

[1036] "Information on communication networks" refers to information sources such as documents, articles, and databases that exist on the internet and other digital networks, and involves using these sources to verify and confirm information.

[1037] A "primary source" refers to a reliable source that provides the origin of information or original data, and serves as a standard for verifying the accuracy of information.

[1038] An "information processing model" refers to an algorithm or program used to analyze information and generate related terms, making the content of the information easier to understand.

[1039] "External information sources" refer to information providers or databases that exist outside the system and are used to verify and confirm information.

[1040] "Means of analyzing search results" refers to methods and devices for evaluating information obtained from external sources and determining the reliability of that information.

[1041] This invention is a system for evaluating the reliability of information, with a server at its core. The server receives information input from the user and performs a series of processes to evaluate the reliability of that information.

[1042] First, the server analyzes the input information using a natural language processing library. Specifically, it uses natural language processing software such as "spaCy" to tokenize the information and extract important elements such as nouns and verbs. This analysis allows the server to identify the subject of the information and related words.

[1043] Next, the server uses a generative AI model to generate relevant keywords from the extracted information. Based on these generated keywords, the server searches external information sources such as the "Google News API" and the "Bing Search API" to collect data to verify the reliability of the information.

[1044] The server analyzes search results obtained from external sources and determines the reliability of the information. Specifically, it compares search results and evaluates whether the information can be verified by other reliable sources. This evaluation result is provided to the user to help them quickly and accurately determine the reliability of the information.

[1045] For example, if a user enters "an article about the announcement of a new technology," the server analyzes the article and extracts keywords such as "technology" and "announcement." Next, it uses these keywords to search external sources and retrieve relevant news articles and official announcements. Finally, the server analyzes the search results and reports to the user whether the content of the article could be verified by other reliable sources.

[1046] An example of a prompt message is, "Please check if the information in this article can be verified by other reliable sources." This prompt allows users to efficiently evaluate the reliability of information.

[1047] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1048] Step 1:

[1049] The user inputs information they want to evaluate the reliability of from their terminal into the system. The server receives this input information and tokenizes the text using the natural language processing library "spaCy". Specifically, it breaks down the information into words and phrases and extracts important elements such as nouns and verbs. This process identifies the subject and related words of the information. The input is information from the user, and the output is the extracted keywords.

[1050] Step 2:

[1051] The server uses a generative AI model to generate relevant keywords based on the keywords extracted in Step 1. These generated keywords are appropriately selected based on the content of the information. Specifically, the generative AI model understands the context of the information and adds highly relevant phrases. This process forms a foundation for a deeper understanding of the information's content. The input is the extracted keywords, and the output is the generated relevant keywords.

[1052] Step 3:

[1053] The server uses the keywords generated in step 2 to search external information sources. Specifically, it uses APIs such as "Google News API" and "Bing Search API" to retrieve relevant news articles and official announcements. This search collects data to verify the reliability of the information. The input is the generated related keywords, and the output is the search results obtained from the external information sources.

[1054] Step 4:

[1055] The server analyzes the search results obtained in step 3 and determines the reliability of the information. Specifically, it compares the search results and evaluates whether the information can be verified by other reliable sources. This evaluation determines the accuracy of the information and reports it to the user. The input is the search results from external sources, and the output is the evaluation result regarding the reliability of the information.

[1056] Step 5:

[1057] The server provides the user with the evaluation results obtained in step 4. Specifically, it notifies the user of its judgment on whether the information is reliable and, if necessary, provides detailed information sources. This notification allows the user to quickly and accurately determine the reliability of the information. The input is the evaluation result regarding the reliability of the information, and the output is the notification to the user.

[1058] (Application Example 2)

[1059] Next, we will describe application example 2 of form example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1060] In modern society, the internet contains a vast amount of news articles, some of which contain unreliable or misleading information. Making decisions based on such information can have significant consequences for individuals and society as a whole. Therefore, there is a need to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[1061] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1062] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information resources on an information network, means for fact-checking, means for comparing with expert opinions, means for notifying the user of the evaluation results, means for analyzing the article content using a generative AI model, and means for generating prompt sentences and comparing them with the information sources. This makes it possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[1063] A "news article" is a collection of information provided on the internet or in print, and includes reports on specific events or incidents.

[1064] "Reliability" is an indicator that shows that information is accurate and free from errors.

[1065] An "information network" is a system for acquiring, transmitting, and sharing information through the internet and other digital communication methods.

[1066] "Information resources" refer to databases, official announcements, and other reliable sources used to assess the credibility of news articles.

[1067] A "primary source" refers to an institution or organization that provides official announcements or data directly related to the events being reported.

[1068] "Fact-checking" is the process of verifying whether reported information is accurate by comparing it with other reliable sources.

[1069] "Expert opinion" refers to evaluations or comments provided by individuals or organizations with knowledge or experience in a particular field.

[1070] A "generative AI model" is an algorithm or system that uses artificial intelligence technology to generate and analyze text.

[1071] A "prompt statement" is an instruction given to a generative AI model, containing instructions for performing a specific task.

[1072] "User" refers to an individual or organization that receives the results of a reliability assessment of a news article.

[1073] The system for implementing this invention uses a server and a user terminal to evaluate the reliability of news articles. The server analyzes the text of the news article and uses information resources on the information network to verify its match with the primary source. Specifically, the server analyzes the article content using a generative AI model, generates prompt sentences, and compares them with the source. This process makes it possible to evaluate the reliability of the article.

[1074] The server notifies the user terminal of the evaluation results. The user terminal receives the evaluation results and displays them to the user. This allows the user to quickly verify the reliability of the news article.

[1075] As a concrete example, consider a scenario where a user is reading a news article stating that "the number of people infected with the new virus is rapidly increasing." The server extracts information about the number of infected people from the article and compares it with official announcements from the Ministry of Health, Labour and Welfare and reports from other reliable media outlets. Based on the comparison results, the server evaluates the reliability of the article as "high," "medium," or "low" and notifies the user's terminal.

[1076] Examples of prompts to input into a generative AI model include the following:

[1077] "Please compare the content of the following news article with reliable sources and assess its reliability. Article content: Article text"

[1078] In this way, by using a server and user terminals, it is possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[1079] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1080] Step 1:

[1081] A user views a news article. The user's device sends the text data of the news article being viewed to the server. The input is the text data of the news article, and the output is the transmission to the server.

[1082] Step 2:

[1083] The server analyzes the text data of the received news articles. Using a generative AI model, it extracts important facts and information from the articles. The input is the text data of the news articles, and the output is the extracted facts and information.

[1084] Step 3:

[1085] The server generates a prompt message based on the extracted facts and information. The generated prompt message includes instructions for cross-referencing with reliable sources. The input is the extracted facts and information, and the output is the generated prompt message.

[1086] Step 4:

[1087] The server uses the generated prompt message to search for information resources on the information network and verify a match with the primary source. The input is the generated prompt message, and the output is the matching result.

[1088] Step 5:

[1089] The server evaluates the reliability of news articles based on the matching results. The evaluation is categorized as "high," "medium," or "low." The input is the matching results, and the output is the reliability evaluation result.

[1090] Step 6:

[1091] The server notifies the user terminal of the evaluation results. The user terminal displays the received evaluation results to the user. The input is the reliability evaluation result, and the output is the notification to the user.

[1092] In this way, it is possible to quickly and accurately evaluate the reliability of news articles and provide this information to users.

[1093] (Example 3)

[1094] Next, we will describe Embodiment 3 of Embodiment Example 3. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1095] When evaluating the reliability of an informational article, it is necessary to efficiently check how well the article's content aligns with expert opinions and primary sources. However, traditional methods have the drawback of requiring recipients of information to spend a lot of time and effort judging the reliability of an article.

[1096] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 3 is realized by the following means.

[1097] In this invention, the server includes means for verifying the consistency of the information article text with primary sources using information from materials on an information network, means for verifying facts, and means for comparing with expert opinions. This makes it possible to quickly and accurately evaluate the reliability of the information article.

[1098] An "information article" is a piece of text, such as news or reports, provided on an information network.

[1099] "Reliability" refers to the characteristic that indicates information is accurate and free from errors and biases.

[1100] An "information network" is a system for acquiring, transmitting, and sharing information via the internet or other means.

[1101] A "primary source" is the origin of information and refers to documents or evidence that provide direct data or facts.

[1102] "Expert opinion" refers to views or comments provided by individuals with advanced knowledge and experience in a particular field.

[1103] "Natural language processing technology" refers to technologies that enable computers to understand, interpret, and generate human language.

[1104] A "data set" is a collection of data gathered for a specific purpose.

[1105] "Evaluation results" refer to information obtained as a result of evaluating the reliability and accuracy of the content of an informational article.

[1106] An "information terminal" is a device used for inputting, displaying, and processing information, and includes computers, smartphones, and other similar devices.

[1107] The following system configuration is used as an embodiment for carrying out this invention.

[1108] The server runs a program to evaluate the reliability of informational articles. This program analyzes the content of the informational articles using natural language processing techniques. Specifically, it uses a generative AI model to tokenize the content of the articles and extract important technical terms and claims. The server then compares the analyzed content of the informational articles with a data set containing expert opinions. This data set includes opinions from experts in various fields and is obtained from materials on the information network.

[1109] The terminal is responsible for sending informational article data entered by the user to the server. Users request analysis from the system by entering the URL or text of the article they want to analyze into the terminal. For example, if a user enters the prompt "I want to check opinions on the latest medical research," the terminal will send that information to the server.

[1110] The server generates and returns to the information terminal the results of its assessment of the reliability of the informational article. The assessment results include information about the article's reliability and its degree of agreement with expert opinions. Users can review these results through their terminal and determine the reliability of the informational article.

[1111] This system allows users to quickly and accurately verify how well the content of an informational article matches expert opinions and primary sources. The specific processing flow in Example 3 will be explained using Figure 15.

[1112] Step 1:

[1113] The user enters the URL or text of the information article they want to analyze into the terminal. The entered data is saved on the terminal as a prompt message. For example, the user might enter the prompt message, "I want to see opinions on the latest medical research."

[1114] Step 2:

[1115] The terminal sends the prompt text entered by the user to the server. The data sent includes the URL and text of the informational article, which serves as input data for the server to perform analysis.

[1116] Step 3:

[1117] The server analyzes the content of the informational article using natural language processing techniques based on the received prompt message. Specifically, it uses a generative AI model to tokenize the article's content and extract important technical terms and arguments. This process outputs the article's content as structured data.

[1118] Step 4:

[1119] The server compares the content of the analyzed informational article with a data set containing expert opinions. This data set includes opinions from experts in various fields and is obtained from materials on the information network. The degree of agreement between the article's claims and the expert opinions is then evaluated.

[1120] Step 5:

[1121] The server generates evaluation results and sends them back to the information terminal. These results include information about the article's reliability and its degree of agreement with expert opinions. This result becomes the output data from the server to the terminal.

[1122] Step 6:

[1123] The terminal displays the evaluation results received from the server to the user. Through the terminal, the user can check the reliability of the article and determine how well the content of the informational article matches expert opinions and primary sources.

[1124] (Application Example 3)

[1125] Next, we will describe application example 3 of form example 3. In the following description, the data processing device 12 will be referred to as a "server," and the headset-type terminal 314 will be referred to as a "terminal."

[1126] While the internet contains a vast amount of news articles and information, it also includes misinformation and unreliable sources. This can lead readers and information recipients to make incorrect judgments, highlighting the need to efficiently evaluate the reliability of news articles and provide accurate information.

[1127] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means.

[1128] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, means for comparing with expert comments, means for analyzing the content of the article using natural language processing technology, and means for comparing with expert opinions and reliable sources using a generative AI model. This makes it possible to evaluate the reliability of news articles with high accuracy and provide accurate information to readers and information recipients.

[1129] A "news article" is a collection of information provided on the internet or in print media, and includes reports about specific events or incidents.

[1130] "Reliability" is an indicator that shows that information is accurate and free from errors.

[1131] A "primary source" refers to the original source or data from which information was first disseminated.

[1132] "Fact-checking" is the process of verifying whether the information provided is accurate.

[1133] "Expert comments" refer to opinions or views expressed by individuals with specialized knowledge in a particular field.

[1134] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[1135] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new information or data.

[1136] "Comparison" is the act of comparing two or more pieces of information or data and evaluating their similarities and differences.

[1137] The system for implementing this invention uses a server and a user terminal to evaluate the reliability of news articles. The server compares the text of the news article with literature information on the internet and verifies its consistency with primary sources. Furthermore, it performs fact-checking and cross-references it with expert comments. This makes it possible to evaluate the reliability of news articles with high accuracy.

[1138] The server analyzes the content of articles using natural language processing techniques. Specifically, it uses a natural language processing library (e.g., spaCy) to extract keywords and arguments from the articles. Next, it uses a generative AI model (e.g., OpenAI's GPT-3) to compare the extracted information with expert opinions and reliable sources. This comparison is used to evaluate the reliability of the articles.

[1139] The user's terminal receives evaluation results provided by the server and displays them to the user. This allows the user to quickly check the reliability of an article and reduces the risk of being misled by misinformation.

[1140] As a concrete example, if a user is reading an article about the effectiveness of a new vaccine, the server analyzes the claims in the article and inputs a prompt message into an AI model asking, "How do medical experts evaluate the effectiveness of this vaccine?" The system then compares the expert opinions returned by the model with the content of the article and evaluates their reliability.

[1141] The flow of the specific processing in Application Example 3 will be explained using Figure 16.

[1142] Step 1:

[1143] A user views a news article. The user's device sends the text data of the news article being viewed to the server. The input is the text data of the news article, and the output is the transmission of data to the server.

[1144] Step 2:

[1145] The server analyzes the text data of received news articles using natural language processing techniques. Specifically, it uses a natural language processing library (e.g., spaCy) to extract keywords and main arguments from the articles. The input is the text data of the news articles, and the output is the extracted keywords and main arguments.

[1146] Step 3:

[1147] The server inputs prompt sentences into a generative AI model (e.g., OpenAI's GPT-3) based on the extracted keywords and claims. These prompt sentences are in the form of seeking expert opinions related to the article's content. The input consists of extracted keywords and claims, and the output is the generated prompt sentences.

[1148] Step 4:

[1149] The generative AI model generates expert opinions and information from reliable sources based on the input prompt sentence. The input is the prompt sentence, and the output is the generated expert opinion or information.

[1150] Step 5:

[1151] The server compares expert opinions and information obtained from the generated AI model with the content of the news article. The comparison evaluates the reliability of the article. The input is the content of the news article and the generated expert opinions, and the output is the reliability evaluation result.

[1152] Step 6:

[1153] The server sends the reliability evaluation results to the user terminal. The user terminal displays the received evaluation results to the user. The input is the reliability evaluation results, and the output is the display to the user.

[1154] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1155] "Example of form 1"

[1156] One embodiment of the present invention provides a news article reliability evaluation system incorporating an emotion engine. This system includes means for verifying the consistency of news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes an emotion engine that recognizes the user's emotions. Specifically, it estimates emotions from the user's facial expressions and tone of voice while reading the article, as well as keystrokes when typing text, and reflects the results in the reliability evaluation. For example, if the user shows anger or distrust, the system lowers the reliability of the article. Conversely, if the user shows satisfaction or trust, the system increases the reliability of the article.

[1157] "Example of form 2"

[1158] Another embodiment of the present invention provides a news article reliability evaluation system incorporating an emotion engine. This system includes means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes means for providing the evaluation results to the article's readers and an emotion engine that recognizes the user's emotions. Specifically, the emotion engine recognizes the emotions the user feels when reading the article and adjusts the reliability evaluation based on those emotions. For example, if the user feels joy or excitement while reading the article, the reliability of that article is evaluated as high.

[1159] "Example of form 3"

[1160] As a further embodiment of the present invention, a news article reliability evaluation system incorporating an emotion engine is provided. This system includes means for verifying the consistency of the news article text with primary sources using information from online literature, means for fact-checking, and means for comparing it with expert comments. Furthermore, it includes means for providing the evaluation results to the article writer and an emotion engine that recognizes the user's emotions. Specifically, the emotion engine recognizes the emotions the article writer felt when writing the article and adjusts the reliability evaluation based on those emotions. For example, if the writer felt anxiety or tension when writing the article, the reliability of that article will be evaluated as low.

[1161] The following describes the processing flow for each example of the form.

[1162] "Example of form 1"

[1163] Step 1: The user selects a news article.

[1164] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[1165] Step 3: The system verifies the facts.

[1166] Step 4: The system compares the information with expert comments.

[1167] Step 5: The emotion engine recognizes the user's emotions.

[1168] Step 6: The system incorporates the results of the emotion engine into the reliability assessment.

[1169] "Example of form 2"

[1170] Step 1: The user selects a news article.

[1171] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[1172] Step 3: The system verifies the facts.

[1173] Step 4: The system compares the information with expert comments.

[1174] Step 5: The emotion engine recognizes the user's emotions.

[1175] Step 6: The system incorporates the results of the emotion engine into the reliability assessment and provides the assessment results to the article's readers.

[1176] "Example of form 3"

[1177] Step 1: The writer creates the news article.

[1178] Step 2: The system uses information from online literature to verify its consistency with primary sources.

[1179] Step 3: The system verifies the facts.

[1180] Step 4: The system compares the information with expert comments.

[1181] Step 5: The emotion engine recognizes the emotions of the article's writer.

[1182] Step 6: The system incorporates the results of the emotion engine into the reliability assessment and provides the assessment results to the article writer.

[1183] (Example 1)

[1184] Next, we will describe Embodiment 1 of Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1185] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, with the vast amount of information available on the internet, verifying whether an article's content aligns with primary sources is not easy. Furthermore, while considering readers' sentiments and expert opinions is necessary when evaluating article reliability, there is a lack of efficient means to do so.

[1186] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1187] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, and means for recognizing user sentiment and reflecting the results in reliability evaluation. This makes it possible to evaluate the reliability of news articles from multiple perspectives and provide accurate evaluation results.

[1188] A "news article" refers to written text or news content created to convey information, and is particularly widely distributed through the internet and print media.

[1189] "Reliability" is an indicator that shows that information or data is accurate and error-free, and it represents the degree to which the recipient of the information can trust its content.

[1190] An "information network" refers to a communication infrastructure for sending and receiving digital data, including the internet, and is a system that allows access to various information sources.

[1191] A "primary source" refers to the place where specific information or data first originated, or the entity that directly provides that information, and forms the basis of reliable information.

[1192] "Fact-checking" is the process of verifying whether the information or data provided matches the actual facts, and is carried out to guarantee the accuracy of the information.

[1193] "Expert opinions" refer to views and analyses provided by individuals or organizations with advanced knowledge and experience in a particular field, and serve as a reference when evaluating the reliability of information.

[1194] "User emotions" refer to the psychological reactions and feelings that users exhibit when reading news articles, and are inferred from their facial expressions, tone of voice, input behavior, etc.

[1195] "Reliability assessment" is an evaluation process used to determine how accurate and reliable the content of a news article is, and it is carried out by considering various factors.

[1196] This invention is a system for evaluating the reliability of news articles, in which the server, terminal, and user elements work together.

[1197] When the server receives a news article, it analyzes the article text using natural language processing techniques. Specifically, the server tokenizes the text and extracts important keywords and numerical data. Common natural language processing libraries can be used for this analysis. Next, the server uses databases on the information network to compare the extracted data with primary sources. Databases such as Google Scholar and PubMed are used to search for information that matches the article's content and calculate the degree of match.

[1198] The device senses the user's facial expressions, tone of voice, and keystrokes when they are typing text while reading news articles. This uses hardware such as a camera, microphone, and keyboard. The device collects this data in real time and estimates the user's emotions. The emotion engine analyzes this emotion data and sends the emotions the user is expressing to the server.

[1199] The server adjusts the reliability rating of news articles based on sentiment data received from the device. If the user expresses anger or distrust, the server will rate the article's reliability lower. Conversely, if the user expresses satisfaction or trust, the server will rate the article's reliability higher.

[1200] As a concrete example, consider a scenario where a user is reading a news article about the effectiveness of a vaccine for the novel coronavirus. The server searches the information network for the data "95% vaccine effectiveness" found in the article and checks if it matches the primary source. Simultaneously, the terminal senses the user's facial expressions and tone of voice, and if the user shows distrust, the server lowers the reliability of the article.

[1201] An example of a prompt to input into a generative AI model is: "Explain how to evaluate the reliability of a news article by comparing the data in the article with primary sources on the information network and adjusting the reliability by considering user sentiment data." This prompt allows the generative AI model to generate a detailed explanation of the system's process.

[1202] The flow of the specific processing in Example 1 will be explained using Figure 17.

[1203] Step 1:

[1204] The server receives news articles from users. The input is the text of the news articles. The server analyzes the article text using natural language processing techniques and extracts important keywords and numerical data. Specifically, it tokenizes the text and identifies parts of speech such as nouns and verbs. The output is a list of the extracted keywords and data.

[1205] Step 2:

[1206] The server searches databases on the information network based on the data extracted in Step 1. Keywords and a list of data are used as input. The server uses databases such as Google Scholar and PubMed to find information that matches the primary sources. Specifically, it sends queries to the databases via APIs to retrieve relevant literature and data. A score indicating the degree of matching is generated as output.

[1207] Step 3:

[1208] The device collects user emotion data when the user reads news articles. Input includes the user's facial expressions, voice tone, and keystrokes during text input. The device uses its camera, microphone, and keyboard to sense and collect this data in real time. Specifically, it uses facial recognition and voice analysis technologies to estimate the user's emotions. The estimated emotion data is then generated as output.

[1209] Step 4:

[1210] The server integrates the similarity score obtained in step 2 with the sentiment data obtained in step 3 to evaluate the reliability of the news article. The similarity score and sentiment data are used as input. The server uses a sentiment engine to calculate the impact of the user's emotions on the reliability evaluation. Specifically, if anger or distrust is indicated, the reliability is rated low, and if satisfaction or trust is indicated, the reliability is rated high. The final reliability evaluation score is generated as output.

[1211] Step 5:

[1212] The server presents the final reliability rating score to the user. The reliability rating score is used as input. The server displays the evaluation results visually, making them easy for the user to understand. Specifically, the reliability score is shown numerically and graphically, providing the user with information to judge the reliability of the article. A visualized evaluation result is generated as output.

[1213] (Application Example 1)

[1214] Next, we will describe Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1215] In modern society, where the reliability of information is paramount, there is a need to quickly and accurately evaluate the reliability of news articles. However, conventional methods are time-consuming in verifying the accuracy of primary sources and fact-checking, and they do not take into account the emotions of users in their reliability assessments. As a result, it is difficult to provide information that is truly reliable to recipients.

[1216] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1217] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, and means for recognizing user sentiment and reflecting the results in the reliability evaluation. This enables rapid and accurate evaluation of the reliability of news articles and personalized reliability evaluations that take user sentiment into consideration.

[1218] A "news article" is a piece of writing intended to convey information, and in particular, it includes content related to current events and social issues.

[1219] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[1220] An "information network" refers to a communication system for sending and receiving digital data, including the internet.

[1221] A "primary source" refers to the source of information or documents and data that provide direct evidence.

[1222] "Fact-checking" refers to the process of verifying whether information is accurate.

[1223] "Expert opinion" refers to the views of individuals who possess advanced knowledge and experience in a particular field.

[1224] "User" refers to an individual or group that uses the system or service.

[1225] "Recognizing emotions" refers to the process of estimating a user's emotional state from their facial expressions, tone of voice, and behavior.

[1226] "Reliability assessment" refers to the process of demonstrating the accuracy and reliability of information using numerical values ​​and indicators.

[1227] "Personalization" refers to adjusting services and information according to the individual characteristics and preferences of each user.

[1228] The system for implementing this invention involves a server and a terminal working together to evaluate the reliability of news articles. The server receives the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. This verifies the facts of the article and compares them with expert opinions.

[1229] When a user views a news article, the device uses its camera and microphone to capture the user's facial expressions and tone of voice. This allows the device to recognize emotions and send the results to a server. The server then adjusts its reliability assessment based on the received emotion data and calculates a final reliability score.

[1230] This system will be implemented as an application installed on devices such as smartphones and smart glasses. Specifically, it will perform natural language processing and sentiment analysis using software such as Python and TensorFlow. spaCy will be used as the natural language processing library, and NLTK will be used as the sentiment analysis library.

[1231] For example, when a user is reading a news article through smart glasses, the article's reliability score is displayed in their field of vision. If the user frowns, the emotion engine detects the distrust and adjusts the reliability score.

[1232] An example of a prompt to input into a generative AI model is: "Please rate the reliability of this news article. The article content is as follows: 'Article Content'. The user's sentiment indicates distrust."

[1233] The flow of a specific process in Application Example 1 will be explained using Figure 18.

[1234] Step 1:

[1235] The device uses its camera and microphone to capture the user's facial expressions and voice tone when they view news articles. This data is used as input for sentiment analysis.

[1236] Step 2:

[1237] The device inputs the acquired facial and voice data into an emotion analysis library (such as NLTK) to estimate the user's emotions. The result of the emotion analysis is output, indicating the emotion the user is expressing (e.g., distrust, satisfaction).

[1238] Step 3:

[1239] The terminal sends the text of the news article to the server. The server analyzes the received article using a natural language processing library (such as spaCy) and extracts facts and data from the article.

[1240] Step 4:

[1241] The server compares the extracted facts and data with bibliographic information on the information network and calculates the degree of match with the primary source. This degree of match is output as the basis for reliability evaluation.

[1242] Step 5:

[1243] The server compares the article content with an expert opinion database and evaluates the degree of agreement between the expert's view and the article. This evaluation result is also taken into account in the reliability assessment.

[1244] Step 6:

[1245] The server receives user sentiment data and incorporates it into the reliability assessment. Specifically, if a user expresses distrust, the reliability score is adjusted.

[1246] Step 7:

[1247] The server calculates the final reliability score and sends it to the terminal. The terminal then displays this score visually to the user.

[1248] Step 8:

[1249] Users can check the displayed reliability score to determine the reliability of a news article. This allows users to read articles while considering the accuracy of the information.

[1250] (Example 2)

[1251] Next, we will describe Example 2 of the morphological example. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1252] When evaluating the reliability of news articles, there is a challenge in objectively judging the accuracy and reliability of the information. Furthermore, since the influence of user emotions on the reliability evaluation of an article is not taken into account, the evaluation results may be influenced by the user's subjectivity. In addition, there is a lack of means to visually present the evaluation results, making it difficult for users to intuitively understand the results.

[1253] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1254] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, means for recognizing user sentiment and adjusting reliability evaluation based on that sentiment, and means for visually displaying the evaluation results. This enables objective and intuitive evaluation of the reliability of news articles, and allows for reliability evaluation that takes user sentiment into consideration.

[1255] A "news article" is written or digital content created to report events or information.

[1256] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[1257] An "information network" is a system for acquiring, transmitting, and sharing information through the internet and other digital communication methods.

[1258] A "primary source" refers to the direct origin or source of an event or piece of information.

[1259] "Fact-checking" is the process of verifying whether information is accurate using other reliable sources.

[1260] "Expert opinion" refers to the views of an individual or organization that possesses advanced knowledge and experience in a particular field.

[1261] "User sentiment" refers to the psychological reactions that users exhibit when reading news articles.

[1262] "Visual display methods" refer to methods of presenting information to users in an easy-to-understand manner using graphs, diagrams, numerical data, etc.

[1263] As an embodiment of this invention, a system for evaluating the reliability of news articles is constructed. The server receives news articles and analyzes their content using natural language processing technology. Specifically, it uses a text analysis library (e.g., spaCy or NLTK) to extract important facts and keywords from the articles.

[1264] The server searches for reliable sources of information through its information network based on extracted facts and keywords. Using APIs, it accesses official police announcements and other media databases to verify the facts presented in the articles. It also sends queries to external databases using HTTP requests.

[1265] The device recognizes the user's emotions as they read the article. It uses the camera and microphone to capture the user's facial expressions and voice, which are then analyzed by an emotion engine. The emotion engine uses machine learning models (e.g., TensorFlow or PyTorch) to classify the user's emotions into categories such as "joy," "anxiety," and "excitement."

[1266] The server integrates fact-checking results with user sentiment assessment results to evaluate the reliability of news articles. If a user expresses positive emotions after reading an article, the reliability score increases. Conversely, if a user expresses negative emotions, more emphasis is placed on cross-referencing with other sources.

[1267] Users receive the reliability evaluation results through their device. The device visually displays the evaluation results, providing the article's reliability score and comments. The evaluation results are presented in graphs and numerical values, in a way that is easy for users to understand.

[1268] As a concrete example, when a user enters a news article about the spread of a new virus, the server analyzes the article's content and verifies relevant facts using external sources. The device captures the user's facial expressions with its camera and analyzes them with an emotion engine. Finally, the reliability evaluation results are displayed on the device, allowing the user to verify the article's reliability.

[1269] Examples of prompts for a generative AI model:

[1270] "Analyze news articles about the spread of the new virus and evaluate their reliability. Verify the facts in the articles with other reliable sources and adjust your evaluation considering user sentiment."

[1271] The flow of the specific processing in Example 2 will be explained using Figure 19.

[1272] Step 1:

[1273] The server receives news articles from users as input. The server analyzes the article content using natural language processing techniques, extracting important facts and keywords. This process utilizes text analysis libraries (e.g., spaCy or NLTK) to analyze the article's grammatical structure and identify keywords. The server then generates a list of extracted keywords and facts as output.

[1274] Step 2:

[1275] The server uses the keywords and facts extracted in Step 1 as input to search for reliable sources of information through the information network. The server accesses external databases using APIs to verify whether the facts in the articles can be confirmed. Specifically, it sends queries to external databases using HTTP requests and retrieves the matching results as output.

[1276] Step 3:

[1277] The device recognizes the user's emotions as input while reading news articles. The device uses its camera and microphone to capture the user's facial expressions and voice, which are then analyzed by an emotion engine. The emotion engine uses machine learning models (e.g., TensorFlow or PyTorch) to classify the user's emotions into categories such as "joy," "anxiety," and "excitement." As output, it generates user emotion data.

[1278] Step 4:

[1279] The server integrates the matching results from Step 2 and the sentiment data from Step 3 as input to evaluate the reliability of the news article. If the user expresses positive sentiment after reading the article, the server increases the reliability score. Conversely, if the user expresses negative sentiment, it places more emphasis on the matching results with other sources. The server generates a reliability evaluation score as output.

[1280] Step 5:

[1281] The user receives the reliability evaluation results through their device. The device visually displays the evaluation results and provides the article's reliability score and comments. Specifically, it presents the evaluation results in graphs and numerical values ​​in a way that is easy for the user to understand. As output, it generates a visually displayed reliability evaluation result.

[1282] (Application Example 2)

[1283] Next, we will describe application example 2 of form example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1284] When evaluating the reliability of news articles, it is necessary to consider not only the content but also the reader's emotions. However, conventional systems do not perform reliability evaluations that take emotions into account, which presents a challenge in providing readers with reliable information.

[1285] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1286] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, and means for recognizing the user's emotions and adjusting the reliability evaluation based on those emotions. This makes it possible to evaluate the reliability of news articles while taking the reader's emotions into consideration and to provide highly reliable information.

[1287] A "news article" refers to written text or news reports created to convey information.

[1288] "Reliability" refers to the degree to which information is judged to be accurate and free from errors.

[1289] An "information network" refers to a communication network used to transmit information, including the internet.

[1290] A "primary source" is a source that provides the origin of information or direct evidence.

[1291] "Fact-checking" is the process of verifying whether information is accurate.

[1292] "Expert comments" refer to opinions or explanations provided by individuals with specialized knowledge in a particular field.

[1293] "Users" refer to individuals or organizations that use a system or service.

[1294] "Emotion" refers to a person's psychological reaction or state, and includes feelings such as joy and sadness.

[1295] "Evaluation" refers to the act of judging value or performance based on specific criteria.

[1296] A "server" is a computer system that provides information and services over a network.

[1297] The system for carrying out this invention includes a program for evaluating the reliability of news articles. The server analyzes the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. This is done using natural language processing libraries (e.g., NLTK, spaCy). Furthermore, it uses external APIs (e.g., Google Fact Check Tools API) to verify the accuracy of the article for fact-checking.

[1298] The device acquires data through its camera and microphone to recognize the user's emotions and analyzes it using an emotion recognition library (e.g., Affectiva). This allows for adjusting the reliability evaluation based on the user's emotions. The evaluation results are displayed on the device's screen and provided to the user.

[1299] As a concrete example, when a user opens a news app on their smartphone and begins reading an article, the server evaluates the article's reliability and displays a reliability score on the screen. If the user smiles while reading the article, the device recognizes that emotion and adjusts the reliability score accordingly.

[1300] An example of a prompt to input into a generative AI model is: "Please evaluate the reliability of this news article. Compare the article's content with external sources and calculate a reliability score, taking into account the user's sentiment."

[1301] The flow of a specific process in Application Example 2 will be explained using Figure 20.

[1302] Step 1:

[1303] The server retrieves the text of the news article selected by the user. It receives the URL or ID of the news article as input and extracts the text data of the article. It generates the text data of the news article to be analyzed as output.

[1304] Step 2:

[1305] The server analyzes the text data of news articles using natural language processing libraries (e.g., NLTK, spaCy). It receives the text data of news articles as input and extracts keywords and important phrases from the articles. As output, it generates data that summarizes the content of the articles.

[1306] Step 3:

[1307] The server uses external APIs (e.g., Google Fact Check Tools API) to fact-check the content of news articles. It receives summary data as input and compares it with external sources. As output, it generates initial assessment data regarding the reliability of the article.

[1308] Step 4:

[1309] The device acquires data on the user's facial expressions and voice through its camera and microphone in order to recognize the user's emotions. It receives real-time video and audio data of the user as input and analyzes it using an emotion recognition library (e.g., Affectiva). As output, it generates data indicating the user's emotional state.

[1310] Step 5:

[1311] The server adjusts the reliability rating of news articles based on the user's emotional state data. It receives initial evaluation data and emotional state data as input and recalculates the reliability score. It generates the adjusted reliability score as output.

[1312] Step 6:

[1313] The device displays the adjusted reliability score to the user. It receives the adjusted reliability score as input and displays it visually on the user's screen. As output, it provides a reliability evaluation result that the user can visually confirm.

[1314] (Example 3)

[1315] Next, we will describe Embodiment 3 of Embodiment Example 3. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1316] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, verifying whether an article's content aligns with primary sources and expert opinions is not easy. Furthermore, it is necessary to consider the impact of the writer's emotions on the article's reliability. An effective system is needed to address these challenges.

[1317] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 3 is realized by the following means.

[1318] In this invention, the server includes means for analyzing information, means for confirming matching with information sources, and means for verifying facts. This makes it possible to evaluate the reliability of news articles from multiple perspectives and to quickly and accurately determine their reliability.

[1319] "Means of analyzing information" refers to a function that uses natural language processing technology to analyze the content of news articles and extract expert arguments and keywords.

[1320] "Means of verifying consistency with information sources" refers to the function of searching relevant literature and databases and comparing their contents to confirm whether the claims in a news article are consistent with primary sources.

[1321] "Means of fact-checking" refers to the function of verifying the accuracy of information contained in news articles by referring to relevant data and expert opinions.

[1322] "Means of comparing with expert opinions" refers to a function that compares and evaluates the claims made in news articles with the opinions and comments of experts in that field.

[1323] "Means of recognizing emotions" refers to the function of analyzing the emotions of news article writers and evaluating the impact of those emotions on the article's credibility.

[1324] "Means for adjusting evaluations" refers to a function that adjusts the reliability evaluation of news articles based on collected information and sentiment analysis results, and generates a final evaluation result.

[1325] To implement this invention, it is necessary to build a system for evaluating the reliability of news articles. The user inputs the news article they wish to evaluate into the system. The server analyzes the content of the article using natural language processing techniques and extracts expert claims and keywords. This analysis uses the Python natural language processing library.

[1326] Next, the server gathers relevant literature and expert opinions from the internet based on the extracted claims. Specifically, it uses academic paper database APIs to search for relevant literature and compare it with the claims in the article. Furthermore, the server uses database query techniques to verify that the claims in the article match those of primary sources.

[1327] Furthermore, the server uses an emotion engine to analyze the emotions of the article's writer. Specifically, it uses a Python emotion analysis library to determine whether the writer is feeling anxious or stressed. This emotion information influences the reliability evaluation.

[1328] Finally, the server integrates the collected information and sentiment analysis results to evaluate the reliability of the news article. The evaluation results are displayed visually to the user. For example, the reliability score may be shown in a graph, with a detailed explanation of which factors influenced the evaluation.

[1329] As a concrete example, when a user enters "an article about the effectiveness of a new virus vaccine," the server analyzes the article and extracts claims about vaccine effectiveness. Next, the server searches for relevant medical papers using an academic paper database and evaluates the accuracy of the claims. Simultaneously, an emotion engine analyzes the writer's emotions and incorporates this into the reliability assessment. Finally, the server generates a reliability score and provides it to the user.

[1330] An example of a prompt to the generative AI model might be: "Please verify whether the medical claims in this article are consistent with expert opinions. Also, please evaluate the impact of the writer's emotions on its credibility." The specific processing flow in Example 3 will be explained using Figure 21.

[1331] Step 1:

[1332] The user inputs a news article they want to evaluate into the system. The server receives the input article and analyzes its content using natural language processing techniques. Specifically, it uses a Python natural language processing library to extract expert claims and keywords from the article. As a result of this analysis, a list of the article's claims and keywords is output.

[1333] Step 2:

[1334] Based on the claims extracted in Step 1, the server collects relevant literature and expert opinions from the internet. Specifically, it uses an academic paper database API to search for relevant literature and compares it with the claims in the article. This process takes a list of claims as input and produces a list of relevant literature and expert comments as output.

[1335] Step 3:

[1336] The server verifies whether the claims in the article match those of the primary source. Specifically, it uses database query techniques to identify the primary source and check for content consistency. This step takes a list of claims as input and outputs results indicating whether or not there is a match.

[1337] Step 4:

[1338] The server uses an emotion engine to analyze the emotions of the article's author. Specifically, it uses a Python emotion analysis library to determine whether the author is feeling anxious or stressed. This process takes the article's text as input and outputs the results of the emotion analysis.

[1339] Step 5:

[1340] The server integrates the collected information and sentiment analysis results to evaluate the reliability of the news article. Specifically, it calculates a reliability score and generates an evaluation result. In this step, relevant literature, expert comments, agreement results, and sentiment analysis results are used as input, and the reliability score is obtained as output.

[1341] Step 6:

[1342] The server provides the user with the final evaluation results. Specifically, it visually displays the reliability score and provides a detailed explanation of which factors influenced the evaluation. In this step, the reliability score is used as input, and the evaluation results are displayed to the user as output.

[1343] (Application Example 3)

[1344] Next, we will describe application example 3 of form example 3. In the following description, the data processing device 12 will be referred to as a "server," and the headset-type terminal 314 will be referred to as a "terminal."

[1345] In today's information society, it is crucial to quickly and accurately assess the reliability of news articles. However, verifying whether an article's content aligns with expert opinions and primary sources is not easy, nor is it feasible to consider the impact of the writer's emotions on its reliability. Therefore, there is a need for a system that comprehensively evaluates the reliability of news articles and provides this information to readers and writers.

[1346] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 3 is realized by the following means.

[1347] In this invention, the server includes means for verifying the consistency of news article text with primary sources using information from literature on an information network, means for fact-checking, means for comparing with expert opinions, means for analyzing the sentiment of the article's writer and adjusting the reliability evaluation, and means for displaying the evaluation results on a display device. This makes it possible to comprehensively evaluate the reliability of news articles and provide this information to readers and writers quickly and accurately.

[1348] A "news article" is a piece of writing created to convey information, and includes reporting on specific events or incidents.

[1349] "Reliability" is a concept that refers to the degree to which information is judged to be accurate and free from errors.

[1350] An "information network" is a system that uses digital communication technologies such as the internet to share information.

[1351] A "primary source" refers to the place or person from which specific information or data first originated; it is the original source of the information.

[1352] "Expert opinion" refers to views or comments from individuals who possess advanced knowledge and experience in a particular field.

[1353] "Sentiment analysis" is a technique that extracts emotions from written text or speech and evaluates the type and intensity of those emotions.

[1354] "Evaluation results" refer to the results of an evaluation based on specific criteria, and in this context, it means an evaluation of the reliability of a news article.

[1355] A "display device" is a device used to visually display information, and includes computer monitors and smartphone screens.

[1356] An "information terminal" is an electronic device used to process and display information, and includes smartphones and tablets.

[1357] The term "creator" refers to the person who produced a particular piece of writing or work.

[1358] The system for implementing this invention uses a server and an information terminal to evaluate the reliability of news articles. The server analyzes the text of the news article and uses bibliographic information on the information network to verify its consistency with primary sources. Furthermore, it performs fact-checking and compares it with expert opinions. This makes it possible to comprehensively evaluate the reliability of the article.

[1359] The server uses a sentiment analysis API (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment of the article's author and adjust the reliability rating accordingly. The evaluation results are displayed on the information terminal's display device, allowing users to check them in real time.

[1360] As a concrete example, when a user opens a news app on their smartphone and selects a specific article, the article's reliability score is displayed on the screen. For instance, if the article reports on a "new medical discovery," the system checks whether the content aligns with the opinions of medical experts and calculates a reliability score.

[1361] An example of a prompt for a generative AI model is, "Verify whether the medical claims in this article are consistent with expert opinions and calculate a reliability score." This prompt prompts the server to perform the necessary data processing and reliability assessment.

[1362] The flow of the specific processing in Application Example 3 will be explained using Figure 22.

[1363] Step 1:

[1364] The user launches a news app on their information terminal and selects a specific article. The input is the news article selected by the user, and the output is the text data of that article. The terminal sends this text data to the server.

[1365] Step 2:

[1366] The server analyzes the text data of received news articles and uses bibliographic information on the information network to verify their match with primary sources. The input is the text data of news articles, and the output is the result of matching with primary sources. The server analyzes the text using a natural language processing library (e.g., spaCy) and verifies the match by calling an external fact-checking API.

[1367] Step 3:

[1368] The server verifies the content of news articles and compares them with expert opinions. The input is the text data of the news article, and the output is the results of the fact-checking and the agreement with expert opinions. The server refers to an expert database to check whether the content of the article is consistent with the opinions of experts.

[1369] Step 4:

[1370] The server uses a sentiment analysis API to analyze the writer's sentiment and adjust the reliability rating accordingly. The input is the text data of the news article, and the output is the result of the writer's sentiment analysis. The server calls a sentiment analysis API (e.g., IBM Watson Natural Language Understanding) to evaluate the writer's sentiment.

[1371] Step 5:

[1372] The server integrates these results to calculate the reliability score of the news article. The inputs are the results of matching primary sources, fact-checking, agreement with expert opinions, and sentiment analysis, and the output is the reliability score. The server integrates this data and runs an algorithm to calculate the reliability score.

[1373] Step 6:

[1374] The server sends the calculated reliability score to the information terminal, and the terminal displays the score to the user. The input is the reliability score, and the output is the reliability score displayed on the terminal's display device. The terminal displays the received score on the screen so that the user can verify it.

[1375] (Other examples)

[1376] Since this is the same as the specific processing described in the other embodiments of the first embodiment above, the explanation will be omitted.

[1377] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1378] The data generation model 58 is a form of so-called generative AI (Artificial Intelligence). One example of the data generation model 58 is ChatGPT (Internet Search).<URL: https: / / openai.com / blog / chatgpt> Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1379] Other examples of generative AI include Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are examples.

[1380] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1381] [Fourth Embodiment]

[1382] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1383] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1384] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1385] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1386] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1387] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1388] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1389] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1390] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1391] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1392] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1393] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1394] Next, the identification process performed by the identification processing unit 290 of the data processing device 12 will be described.

[1395] "Example of form 1"

[1396] One embodiment of the present invention is a system for evaluating the reliability of news articles. This system has a means for verifying the consistency between the text of a news article and its primary source using information from online sources. Specifically, it compares the facts and data described in the article with publicly available papers and databases on the internet and calculates the degree of agreement.

[1397] "Example of form 2"

[1398] Furthermore, the system of the present invention also includes means for fact-checking. This verifies whether the facts and information described in the article can be confirmed by other reliable sources. For example, it verifies whether the incident reported in the article has been reported in official police statements or by other media outlets.

[1399] "Example of form 3"

[1400] Furthermore, the system of the present invention includes means for cross-referencing with expert comments. This means that the specialized content and views described in the article are evaluated in comparison with the opinions and comments of experts in that field. For example, it checks whether the medical cl...

Claims

1. Equipped with a processor, The aforementioned processor, Retrieve the full text of the news article, Based on the acquired news article text, a summary of the news article text is created, a prompt incorporating the summary is generated, and the prompt instructs to search for bibliographic information on an information network, identify the primary source of the news article text from the bibliographic information, and evaluate the degree of matching between the identified primary source and the news article text. The generated prompt is input to the generation AI model, and the degree of match between the primary source and the news article text included in the response from the generation AI model is obtained. Furthermore, the system searches for at least one of the facts and data in the news article text from publicly available information sources on the information network, compares the search results obtained from the search with at least one of the facts and data in the news article text to calculate the degree of agreement, calculates a reliability score that quantifies the reliability of the news article text based on the agreement information obtained from the generating AI model and the calculated agreement score, and transmits the reliability score to the user's terminal. system.

2. The system according to claim 1, wherein the processor calculates the degree of match using a text similarity calculation algorithm in the matching and calculation of the reliability score.

Images