System

Abstract

A system is provided.SOLUTION: This system includes a means for collecting data through a network and acquiring the text and metadata of a news article for evaluating the reliability of the news article, a means for analyzing the acquired text and metadata and calculating the reliability score of information, and a means for transmitting the calculated reliability score to a user terminal and displaying a warning.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, 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 documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Describe the "problem that the invention aims to solve" and "means for solving the problem" in the patent specification.

[0005] ---

[0006] In recent years, a large amount of information has been shared via the internet and social media, and this information can contain false or misleading information. Fake news, fabricated videos and images, and fraudulent data collection can have serious impacts on individuals and society as a whole, making it urgent to address these issues. However, existing technologies lack the means to effectively solve these problems, and more advanced detection methods are needed. [Means for solving the problem]

[0007] The present invention relates to a system for ensuring the accuracy of information on the Internet, and includes the following means.

[0008] A system including: a means for collecting data via a network and acquiring the text and metadata of news articles to evaluate the reliability of the news articles; a means for analyzing the acquired text and metadata and calculating a reliability score of the information; and a means for transmitting the calculated reliability score to a user terminal and displaying a warning.

[0009] Furthermore, a system is provided that includes a means for receiving uploaded video or image data to analyze the video or image and detect editing and unauthorized manipulation, a means for analyzing the received data and identifying traces of unauthorized editing, and a means for transmitting the analysis results to a user terminal and displaying a warning.

[0010] The system also includes a means for monitoring users' online activities and collecting site information to monitor social media sites and detect unauthorized data collection and tracking, a means for analyzing the collected site information, determining attempts at unauthorized data collection or tracking, and a means for transmitting the determination results to the user's terminal and displaying a warning. In this way, it is possible to detect and prevent false information and unauthorized activities on the Internet and realize a highly reliable information environment.

[0011] ---

[0012] The "body of a news article" is the main body of text created by a particular news organization or individual to convey information.

[0013] "Metadata" is supplementary information that accompanies data such as news articles, videos, and images, and includes information such as the author, publication date, and title.

[0014] A "credibility score" is a number calculated to determine the credibility of a news article, video, or image, and is an index that indicates the accuracy and reliability of the information.

[0015] "Video data" refers to data containing information in the form of moving images, including continuous images and audio.

[0016] "Image data" means data containing information in the form of a still image, providing a single visual representation.

[0017] "Editing artifacts" are evidence of artificial alterations or manipulation of footage or images, including unnatural color inconsistencies and shadow discontinuities.

[0018] "Unauthorized data collection" refers to the act of acquiring data without the user's consent, including the collection of personal information and behavioral history.

[0019] "Tracking" is a technology that tracks user behavior on websites and applications for the purposes of optimizing advertising and analyzing user behavior.

[0020] A "warning" is a notification message that is provided visually or audibly to inform the user of a risk or problem.

[0021] "Analysis" is the process of examining data in detail to understand its structure and properties, and is done using algorithms and models.

[0022] A "user terminal" is an electronic device for personal use, including a personal computer, a smartphone, a tablet, etc.

[0023] A "network" is a system in which multiple computers and devices are interconnected using the Internet or other communication means.

[0024] A "system" is an entire organization that includes multiple devices, software, and networks combined to achieve a specific function. [Brief explanation of the drawings]

[0025] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14]FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0026] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0027] First, the terms used in the following description will be explained.

[0028] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0029] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0030] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0031] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple 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), etc.

[0032] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0033] [First embodiment]

[0034] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0035] 1, a 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.

[0036] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0037] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0038] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the 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.

[0039] 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 of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0040] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0041] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0042] 2, in the data processing device 12, a specific process 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" according to the technology of the present 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 process 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.

[0043] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0044] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0045] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. 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."

[0046] ---

[0047] The present invention relates to a system for ensuring the accuracy of information on the Internet, and can be implemented as follows: The system of the present invention aims to evaluate and analyze news articles, video and image data, and information through social media to ensure their reliability and accuracy.

[0048] News article credibility assessment

[0049] When a user views a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the news article text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the content of the text, the reliability of the source, past history, etc. The calculated reliability score and analysis results are then sent to the user's device, and a warning message such as "This news article has low reliability" is displayed on the user's device.

[0050] Specific examples

[0051] When a user clicks on a news article shared on a social networking site, the URL is sent to the server, which analyzes the news article and displays a warning message on the user's device if the article has a low reliability score. In this case, the user can confirm the reliability of the news article.

[0052] Detecting edited video and image data

[0053] When a user browses video or image files online, they can upload them to the system. The user's device sends the video or image data to the server, which analyzes it. The server's AI then checks the video or image for signs of editing and identifies unnatural changes, such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message is displayed, such as "This video may be faked."

[0054] Specific examples

[0055] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If any unauthorized editing is detected, a warning is displayed on the user's device.

[0056] Social Media Monitoring

[0057] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed in real time.

[0058] Specific examples

[0059] As users browse social media sites, the browser extension monitors the site's scripts and requests, and if the server determines that the site is engaging in unauthorized data collection, a warning message is displayed on the user's device.

[0060] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection, allowing them to enjoy a safer and more reliable information environment.

[0061] ---

[0062] This concludes the description of the present invention, which provides concrete steps to improve the reliability of information in news articles, video and image data, and social media monitoring.

[0063] The processing flow will be explained below.

[0064] News article credibility assessment

[0065] Step 1:

[0066] A user clicks on a link to a news article. The user's device sends this link information to the server.

[0067] Step 2:

[0068] The server retrieves the HTML data of the news article from the received URL.

[0069] Step 3:

[0070] The server extracts the news article text and metadata from the HTML data it has obtained.

[0071] Step 4:

[0072] The generation AI placed on the server analyzes the text of the news article using natural language processing (NLP) technology and evaluates the article's content, source reliability, past history, etc.

[0073] Step 5:

[0074] The server calculates a reliability score based on the analysis results.

[0075] Step 6:

[0076] The server sends the calculated reliability score and a summary of the analysis results to the user's device.

[0077] Step 7:

[0078] The analysis results received by the user's device are displayed, and if necessary, a warning message such as "This news article is unreliable" is displayed.

[0079] Detecting edited video and image data

[0080] Step 1:

[0081] A user uploads video or image data to the system.

[0082] Step 2:

[0083] The user's terminal transmits the uploaded file data to the server.

[0084] Step 3:

[0085] The server receives the file data.

[0086] Step 4:

[0087] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[0088] Step 5:

[0089] The server sends the analysis results to the user's device.

[0090] Step 6:

[0091] The analysis results received by the user's device are displayed, along with a warning message such as "This video may be faked."

[0092] Social Media Monitoring

[0093] Step 1:

[0094] A user browses to a social media site.

[0095] Step 2:

[0096] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[0097] Step 3:

[0098] The user's terminal transmits the collected site information to the server.

[0099] Step 4:

[0100] The server analyzes the received site information.

[0101] Step 5:

[0102] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[0103] Step 6:

[0104] The server sends the analysis results to the user's device.

[0105] Step 7:

[0106] The analysis results received by the user's device are displayed, and a warning message such as "This site is collecting data illegally" is displayed in real time.

[0107] ---

[0108] These are the specific processing steps for assessing the credibility of news articles, detecting edited video and image data, and monitoring social media. By following these steps, the system can efficiently evaluate information and provide necessary warnings to users.

[0109] Example 1

[0110] Next, a description will be given of 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."

[0111] The problem is that it is difficult for users to easily determine the authenticity of news articles, video and image data, and information on social media, which is flooding the internet. This can lead to the spread of false information and the violation of privacy through the unauthorized collection of data.

[0112] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0113] In this invention, the server includes: means for transmitting a specific URL of a news article from a user's device to the server; means for receiving the transmitted URL at the server and acquiring the text and metadata of the news article; means for analyzing the acquired text and metadata using a generative AI model located on the server to calculate a reliability score of the information; means for transmitting the calculated reliability score and the analysis result to the user's device and displaying them as a warning message; means for receiving video or image data uploaded from the user's device and transmitting the received video or image data to the server to detect editing and unauthorized manipulation; means for analyzing the transmitted data using a generative AI model located on the server to identify traces of unauthorized editing; means for transmitting the analysis result to the user's device and displaying a warning message; means for monitoring the user's online activity using a browser extension and collecting information about the sites accessed; means for transmitting the collected site information to the server, analyzing it using the generative AI model located on the server, and determining attempts of unauthorized data collection or tracking; and means for transmitting the determination result to the user's device and displaying it as a warning message in real time. This allows users to quickly and easily evaluate the reliability of news articles, video and image data, and information on social media and protect themselves from unauthorized data collection.

[0114] A "news article" is an article or news report published on the Internet, including the text and metadata such as the title, author, and publication date.

[0115] "URL" is an abbreviation for Uniform Resource Locator, and is a character string that indicates the address of a specific resource on the Internet.

[0116] A "generative AI model" is an artificial intelligence model used to perform natural language processing and data analysis, such as a machine learning model that generates sentences and performs data analysis.

[0117] The "trustworthiness score" is a numerical representation of the reliability and accuracy of news articles, video and image data, social media information, etc., and is an indicator of their reliability.

[0118] "Analysis" is the process of using generative AI models to examine the content of data in detail and evaluate its content, the reliability of its sources, and evidence of editing.

[0119] A "warning message" is a message that displays important information or a warning to the user, and is particularly a warning about information that is unreliable or inaccurate.

[0120] "Video data" refers to digital data stored in a moving image format, including video clips and movies.

[0121] "Image data" refers to digital data stored in still image format, including photographs and drawings.

[0122] "Editing traces" are unnatural changes or inconsistencies that serve as evidence that video or image data has been edited or manipulated.

[0123] A "browser extension" is additional software that extends the functionality of a web browser and has the ability to monitor a user's online activities and collect specific site information.

[0124] "Unauthorized data collection" refers to the act of illegally collecting personal data or usage information without the user's consent.

[0125] "Tracking" refers to the act of tracking a user's website browsing behavior and collecting and analyzing that data.

[0126] This invention relates to a system that monitors news articles, video and image data, and social media to evaluate the accuracy and reliability of the information. This system uses a generative AI model to evaluate the reliability of news articles, detect editing traces in video and image data, and monitor social media.

[0127] News article credibility assessment

[0128] When a user browses a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the text and metadata of the news article based on the received URL. A generative AI model (e.g., GPT-4) deployed on the server analyzes the text and metadata of the retrieved news article and calculates a reliability score. The analysis result and reliability score are then sent back to the user's device and displayed as a warning message on the user interface.

[0129] Specific examples

[0130] When a user clicks on a news article shared on a social networking site, the URL is sent to a server. The server analyzes the news article and calculates a reliability score using a generative AI model. If the reliability score is low, a warning message stating "This news article is unreliable" is displayed on the user's device.

[0131] Prompt Sentence Examples

[0132] "Analyze the text of news articles and calculate a credibility score."

[0133] Detecting edited video and image data

[0134] When a user views video or image files on the internet, they upload these files to the system. The user's device sends the video or image data to a server, which analyzes the data. A generative AI model (e.g., DALL-E) located on the server checks the video or image for signs of editing and identifies unnatural changes such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message such as "This video may be faked" is displayed.

[0135] Specific examples

[0136] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If unauthorized editing is detected, a warning message is displayed on the user's device stating, "This video may have been forged."

[0137] Prompt Sentence Examples

[0138] "Please detect any editing signs in this image and assess its authenticity."

[0139] Social Media Monitoring

[0140] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether or not unauthorized data collection or tracking is taking place. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is unauthorized data collection" is displayed in real time.

[0141] Specific examples

[0142] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's device will display a warning message stating, "This site is engaging in unauthorized data collection."

[0143] Prompt Sentence Examples

[0144] "Check if this site is engaging in unauthorized data collection or tracking."

[0145] Through this system, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling users to enjoy a safer and more reliable information environment.

[0146] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0147] News article credibility assessment

[0148] Step 1:

[0149] When a user views a news article, they click on the news article's URL.

[0150] Specific actions

[0151] When a user clicks on a news article link in an internet browser, JavaScript runs and retrieves the URL of the news article.

[0152] input

[0153] The URL of the news article the user clicked on.

[0154] output

[0155] The URL of the news article is retrieved on the user's device.

[0156] Step 2:

[0157] The device sends the URL of the news article to the server.

[0158] Specific actions

[0159] The JavaScript code sends the obtained URL to the backend server as a POST request.

[0160] input

[0161] The URL of the news article.

[0162] output

[0163] The URL of the news article is sent to the server.

[0164] Step 3:

[0165] The server retrieves the text and metadata of the news article based on the received URL.

[0166] Specific actions

[0167] The server uses Python's requests library to send requests to news sites, and extracts the text and metadata from the HTML using BeautifulSoup or similar.

[0168] input

[0169] The URL of the news article.

[0170] output

[0171] News article text and metadata.

[0172] Step 4:

[0173] The server sends the acquired news article text and metadata to the generative AI model, which generates an analysis prompt.

[0174] Specific actions

[0175] The server inputs the text and metadata of the news article into a generative AI (e.g., GPT-4) and generates a prompt saying, "Please rate the credibility of this article."

[0176] input

[0177] News article text and metadata.

[0178] output

[0179] The generated analysis prompt.

[0180] Step 5:

[0181] A generative AI model analyzes news articles and calculates a credibility score.

[0182] Specific actions

[0183] The generative AI model analyzes the text and metadata of the input news article and calculates a credibility score.

[0184] input

[0185] Parse prompts, news article body text and metadata.

[0186] output

[0187] Reliability scores and analysis results.

[0188] Step 6:

[0189] The server transmits the calculated reliability score and the analysis result to the user terminal.

[0190] Specific actions

[0191] The server returns the reliability score and analysis results in JSON format to the user device via the REST API set at the endpoint.

[0192] input

[0193] Reliability scores and analysis results.

[0194] output

[0195] The reliability score and analysis results sent to the user's device.

[0196] Step 7:

[0197] Based on the information received by the terminal, a warning message is displayed on the user interface.

[0198] Specific actions

[0199] The user interface (HTML, CSS, JavaScript) receives the reliability score and analysis results from the server and displays a warning message such as "This news article is not reliable" as a popup or banner.

[0200] input

[0201] Reliability scores and analysis results.

[0202] output

[0203] The warning message displayed in the user interface.

[0204] Detecting edited video and image data

[0205] Step 1:

[0206] A user uploads a video or image file onto the Internet.

[0207] Specific actions

[0208] The user interface displays a file selection dialog and the user selects a file.

[0209] input

[0210] Video or image files selected by the user.

[0211] output

[0212] The selected file will be uploaded to the system.

[0213] Step 2:

[0214] The terminal transmits the selected video or image data to the server.

[0215] Specific actions

[0216] Use the JavaScript FormData object to send the selected file to the server via a POST request.

[0217] input

[0218] Video or image files.

[0219] output

[0220] Video or image files sent to the server.

[0221] Step 3:

[0222] The server sends the received video or image data to the generative AI model, which generates an analysis prompt.

[0223] Specific actions

[0224] The server inputs the selected file data into a generation AI (e.g., DALL-E) and generates a prompt saying, "Please detect any editing traces in this image."

[0225] input

[0226] Uploaded video or image files.

[0227] output

[0228] The generated analysis prompt.

[0229] Step 4:

[0230] The generative AI model checks the video or image data for signs of editing and sends the results back to the server.

[0231] Specific actions

[0232] The AI ​​detects unnatural changes, color inconsistencies, and shadow discontinuities in the image and sends the analysis results back to the server.

[0233] input

[0234] Video or image file, analysis prompt.

[0235] output

[0236] Analysis results including traces of editing.

[0237] Step 5:

[0238] The server transmits the analysis results to the user terminal.

[0239] Specific actions

[0240] The server returns the analysis results in JSON format to the user device via the REST API set at the endpoint.

[0241] input

[0242] Analysis results including traces of editing.

[0243] output

[0244] Analysis results sent to the user's device.

[0245] Step 6:

[0246] Based on the analysis results, the device displays a warning message on the user interface.

[0247] Specific actions

[0248] The user interface receives the analysis results and displays a warning message such as "This footage may be faked" as a pop-up or banner.

[0249] input

[0250] Analysis results including traces of editing.

[0251] output

[0252] The warning message displayed in the user interface.

[0253] Social Media Monitoring

[0254] Step 1:

[0255] The browser extension monitors your online activity as you browse social media sites.

[0256] Specific actions

[0257] The browser extension collects the URLs of the sites you visit and the scripts those sites run.

[0258] input

[0259] Information about the social media sites you visit.

[0260] output

[0261] Site Information Collected.

[0262] Step 2:

[0263] The terminal sends the collected site information to the server.

[0264] Specific actions

[0265] The browser extension sends the collected data in JSON format to the server via a POST request.

[0266] input

[0267] Site Information Collected.

[0268] output

[0269] Site information sent to the server.

[0270] Step 3:

[0271] The server analyzes the site information it receives to determine whether unauthorized data collection or tracking is occurring.

[0272] Specific actions

[0273] The server analyzes the collected data against industry-standard security rules (for example, the OWASP threat list).

[0274] input

[0275] Received site information.

[0276] output

[0277] Verification of fraudulent data collection or tracking.

[0278] Step 4:

[0279] The server transmits the determination result to the user terminal.

[0280] Specific actions

[0281] The server returns the result of the judgment in JSON format to the user device via the REST API set at the endpoint.

[0282] input

[0283] Verification of fraudulent data collection or tracking.

[0284] output

[0285] The judgment result sent to the user terminal.

[0286] Step 5:

[0287] Based on the judgment result, the terminal displays a warning message on the user interface.

[0288] Specific actions

[0289] The user interface receives the judgment result and displays a warning message such as "This site is collecting data illegally" as a pop-up or banner.

[0290] input

[0291] Judgment result.

[0292] output

[0293] The warning message displayed in the user interface.

[0294] (Application example 1)

[0295] Next, a description will be given of Application 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."

[0296] The decline in the reliability of information on the Internet and the resulting spread of misinformation and fake news have become a social problem. Furthermore, the increasing number of edited videos and images, as well as the increasing number of illegal data collection and tracking practices on social media, are making it difficult for users to determine the authenticity of information. These problems must be resolved quickly and accurately, and users must be provided with reliable information.

[0297] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0298] In this invention, the server includes means for collecting data via a network and acquiring news article text and metadata, means for analyzing the acquired text and metadata and calculating a reliability score for the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for receiving and analyzing video or image data and detecting editing and unauthorized manipulation, and means for monitoring and determining unauthorized data collection and tracking on social media sites and displaying a warning to the user. This allows users to quickly evaluate the reliability of information on the Internet and be protected from unauthorized data collection.

[0299] A "network" is a collection of computers and communication devices that allow information and data to be sent and received.

[0300] "Data" refers to the information being analyzed, such as news articles, videos, images, and social media information.

[0301] A "news article" is written information such as a news report posted on the Internet.

[0302] "Body" refers to the main content of a news article.

[0303] "Metadata" is supplemental information and attribute data associated with a news article.

[0304] A "trust score" is a numerical assessment of the trustworthiness of a particular article or piece of information.

[0305] A "warning" is a message that is displayed to the user to call attention to the situation.

[0306] "Video" is data that contains moving visual information.

[0307] An "image" is data containing still visual information.

[0308] "Editing evidence" refers to evidence of changes that indicate an image or video has been modified.

[0309] "Social media" is an online platform for sharing information between users.

[0310] "Unauthorized data collection" refers to the illegal act of collecting and using user information without permission.

[0311] "Tracking" is the act of following a user's online activities.

[0312] A "user terminal" is a device used to view information and receive alerts.

[0313] A "generative AI model" is a machine learning model for performing natural language processing and data analysis.

[0314] A "prompt sentence" is a sentence used to input instructions or questions into a generative AI model.

[0315] This invention provides a system for evaluating the reliability of information on the Internet. The system analyzes news articles, video and image data, and information on social media, and conveys the reliability to users.

[0316] Program Generation

[0317] The following describes an outline of a program and its processing for realizing the present invention.

[0318] Hardware / Software used

[0319] Hardware: User's smartphone, server

[0320] Software: Mobile apps, browser extensions, generative AI models (e.g., GPT-4), data analysis tools

[0321] News article credibility assessment

[0322] When a user views a news article on their smartphone, the URL of that news article is automatically sent to a server. A generative AI model (e.g., GPT-4) on the server retrieves and analyzes the news article text and metadata from the URL. Based on the analysis results, a reliability score is calculated and sent to the user's smartphone. If the reliability score is low, the user's smartphone displays a warning message such as "This news article is unreliable."

[0323] Examples:

[0324] When a user clicks on a news article shared on a social networking site, the URL is sent to a server, which analyzes the news article and, if the reliability score is low, displays a warning message on the user's smartphone saying, "This news article is unreliable."

[0325] Prompt for the generative AI model:

[0326] News article URL: [example.com / news123]

[0327] Rate the article's credibility and calculate a credibility score, and explain why it contains exaggerated or unreliable content.

[0328] Detecting traces of editing of video and image data

[0329] When a user views a video or image file on their smartphone, the file is automatically uploaded to a server. A generative AI model on the server analyzes the video or image and detects any signs of editing. If there is a problem with its authenticity, that information is sent to the user's smartphone, warning them that "this video may be faked."

[0330] Examples:

[0331] When a user downloads a video shared on a social networking site and uploads it to the system, the server analyzes the video, and if any fraud is detected, a warning message will appear on the smartphone saying, "This video may be faked."

[0332] Social Media Monitoring

[0333] When a user visits a social media site in their browser, the browser extension monitors the user's online activity and sends the site information to a server. An analysis tool on the server identifies any unauthorized data collection or tracking attempts and sends the results to the user's smartphone. If any unauthorized activity is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed.

[0334] Examples:

[0335] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's smartphone will display a warning message stating, "This site is engaging in unauthorized data collection."

[0336] In this way, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling them to enjoy a safer and more reliable information environment.

[0337] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0338] Program flow:

[0339] News article credibility assessment

[0340] Step 1:

[0341] The user terminal inputs the URL of the news article.

[0342] The entered URL is sent from the user's smartphone to the server.

[0343] Step 2:

[0344] The server retrieves the text and metadata of the news article based on the received URL.

[0345] The server uses a web scraping tool to retrieve data from the URL.

[0346] Step 3:

[0347] The server parses the retrieved text and metadata.

[0348] A generative AI model (e.g., GPT-4) is used to calculate a confidence score for the text.

[0349] Step 4:

[0350] The server transmits the calculated reliability score to the user terminal.

[0351] The reliability score and analysis results are displayed on the user's smartphone.

[0352] Step 5:

[0353] If the user terminal has a low reliability score, a warning message is displayed.

[0354] Users may see a message that reads, "This news article is not reliable."

[0355] Detecting traces of editing of video and image data

[0356] Step 1:

[0357] The user uploads a video or image to the device.

[0358] The uploaded data is automatically sent to the server.

[0359] Step 2:

[0360] The server analyzes the received video or image data.

[0361] Generative AI models are used to detect edits (e.g., unnatural color tones and shadows).

[0362] Step 3:

[0363] The server transmits the analysis results to the user terminal.

[0364] A warning message appears on the user's smartphone saying, "This video may be faked."

[0365] Step 4:

[0366] The user terminal displays a warning message and presents the analysis results.

[0367] Users can see the specific locations and reasons for unauthorized edits.

[0368] Social Media Monitoring

[0369] Step 1:

[0370] A user browses a social media site.

[0371] Browser extensions monitor your online activity and collect site information.

[0372] Step 2:

[0373] The server analyzes the collected site information.

[0374] Generative AI models identify potential fraudulent data collection and tracking.

[0375] Step 3:

[0376] The server transmits the determination result to the user terminal.

[0377] A warning message appears on the user's smartphone saying, "This site is collecting data illegally."

[0378] Step 4:

[0379] The user terminal displays a warning message and provides details of the fraudulent activity.

[0380] Users will be given a detailed explanation of what fraudulent activity is taking place.

[0381] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0382] ---

[0383] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[0384] Evaluating the credibility of news articles and using sentiment engines

[0385] When a user views a news article online, the user's device sends the news article's URL to a server. The server retrieves the news article's text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the text content, source reliability, past history, etc. The user's emotion engine then recognizes the user's emotional state (excited, angry, calm, etc.). The calculated reliability score and analysis results are sent to the user's device, and the way the warning is displayed is adjusted based on the user's emotional state detected by the emotion engine. When a warning message such as "This news article is unreliable" is displayed, if the user is angry, the warning will be displayed more gently, and if the user is calm, a more direct warning will be displayed.

[0386] Specific examples

[0387] When a user clicks on a news article shared on a social networking site, the URL is sent to the server. The server analyzes the news article, and if the reliability score is low, a warning message is displayed on the user's device. At the same time, the emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a gentle warning if the user is excited.

[0388] Detecting edited video and image data and using emotion engines

[0389] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generation AI checks the video or image for signs of editing and identifies any unauthorized changes. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. The warning, "This video may be faked," is adjusted according to the user's state.

[0390] Specific examples

[0391] Users download videos posted on social media and upload them to the system. The server analyzes the video to determine whether there are any signs of unauthorized editing. If unauthorized editing is detected, the emotion engine recognizes the user's emotions and displays a gentle warning message if the user is feeling angry.

[0392] Social media monitoring and sentiment engine utilization

[0393] When a user visits a social media site in a browser, a browser extension is used to monitor the user's online activity. The user's device sends information about the site being accessed to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The result of the determination is sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message is displayed encouraging them to stay calm. The warning, "This site is engaging in unauthorized data collection," is adjusted according to the user's state.

[0394] Specific examples

[0395] As a user browses social media sites, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, an emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a more gentle message if the user is impatient.

[0396] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into account, providing more effective and responsive warnings. This allows users to enjoy a safer and more reliable information environment.

[0397] ---

[0398] This concludes the description of the embodiment of the present invention. This system can more effectively ensure the accuracy of information and provide appropriate warnings to users by evaluating the reliability of news articles, detecting edits in video and image data, and monitoring social media, as well as taking into account the emotional state of the user.

[0399] The processing flow will be explained below.

[0400] Evaluating the credibility of news articles and using sentiment engines

[0401] Step 1:

[0402] A user clicks on a news article link, and the user's device sends the news article URL to the server.

[0403] Step 2:

[0404] The server retrieves the HTML data of the news article from the received URL.

[0405] Step 3:

[0406] The server extracts the news article text and metadata from the HTML data it has obtained.

[0407] Step 4:

[0408] The generation AI placed on the server analyzes the text of news articles using natural language processing (NLP) technology and evaluates the content of the article, the reliability of the source, past history, etc.

[0409] Step 5:

[0410] The server calculates a reliability score based on the analysis results.

[0411] Step 6:

[0412] The server queries the user's emotion engine for the user's most recent emotional state.

[0413] Step 7:

[0414] The user's device activates an emotion engine to detect the user's emotional state (excitement, anger, calmness, etc.).

[0415] Step 8:

[0416] The emotion engine transmits the user's emotional state to the server.

[0417] Step 9:

[0418] The server adjusts how the warning message is displayed depending on the confidence score and the detected emotional state.

[0419] Step 10:

[0420] The server sends a reliability score and a warning message to the user's device.

[0421] Step 11:

[0422] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[0423] Detecting edited video and image data and using emotion engines

[0424] Step 1:

[0425] A user uploads video or image data to the system.

[0426] Step 2:

[0427] The user's terminal transmits the uploaded file data to the server.

[0428] Step 3:

[0429] The server receives the file data.

[0430] Step 4:

[0431] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[0432] Step 5:

[0433] The server creates a warning message based on the analysis results.

[0434] Step 6:

[0435] The server queries the user's emotion engine for the emotional state.

[0436] Step 7:

[0437] The user's terminal activates an emotion engine to detect the user's emotional state.

[0438] Step 8:

[0439] The emotion engine transmits the user's emotional state to the server.

[0440] Step 9:

[0441] The server tailors the warning message based on the analysis results and emotional state.

[0442] Step 10:

[0443] The server sends the analysis results and warning messages to the user's terminal.

[0444] Step 11:

[0445] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[0446] Social media monitoring and sentiment engine utilization

[0447] Step 1:

[0448] A user browses to a social media site.

[0449] Step 2:

[0450] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[0451] Step 3:

[0452] The user's terminal transmits the collected site information to the server.

[0453] Step 4:

[0454] The server analyzes the received site information.

[0455] Step 5:

[0456] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[0457] Step 6:

[0458] The server generates a verdict and a warning message.

[0459] Step 7:

[0460] The server queries the user's emotion engine for the emotional state.

[0461] Step 8:

[0462] The user's terminal activates an emotion engine to detect the user's emotional state.

[0463] Step 9:

[0464] The emotion engine transmits the user's emotional state to the server.

[0465] Step 10:

[0466] The server adjusts the warning message based on the result of the assessment and the emotional state.

[0467] Step 11:

[0468] The server sends the judgment result and a warning message to the user's terminal.

[0469] Step 12:

[0470] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[0471] ---

[0472] In this way, by taking into account the user's emotional state in addition to assessing the reliability of news articles, detecting traces of editing in video and image data, and monitoring social media, we can more effectively ensure the accuracy of information and provide appropriate warnings to users.

[0473] Example 2

[0474] Next, a description will be given of Example 2. 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."

[0475] The internet is flooded with information, often containing unreliable news articles and fraudulently manipulated video and images. Social media sites also sometimes conduct fraudulent data collection and tracking. This information not only causes confusion and anxiety among users, but can also lead to social unrest. Therefore, it is important to evaluate the reliability of this information and provide appropriate warnings to users. However, conventional systems often display uniform warnings without considering the user's emotional state, which can detract from the user experience.

[0476] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0477] In this invention, the server includes means for acquiring news article text and metadata, means for analyzing the acquired text and metadata and using a generative AI model to calculate a reliability score for the information, means for using an emotion engine to recognize the user's emotional state, means for adjusting a warning based on the calculated reliability score and the user's emotional state and sending the adjusted warning message to the user terminal, and means for displaying the adjusted warning message on the user terminal. This makes it possible to evaluate the reliability of news articles, video / image data, and social media site information, and to provide flexible warnings that take the user's emotional state into consideration.

[0478] "Means for obtaining news article text and metadata" refers to a system for collecting the text of news articles published on the Internet and related metadata such as publication date, author information, and URL.

[0479] "Means for using generative AI models" refers to a mechanism for using generative artificial intelligence models to analyze collected data and output the results.

[0480] An "emotion engine for recognizing the user's emotional state" is an engine that can identify and analyze the user's emotions, and is a mechanism for determining whether the user is excited, angry, calm, etc.

[0481] The "means for calculating the reliability score of information" is a system that analyzes the content and metadata of news articles, video and image data, and expresses the reliability of that information as a quantified score.

[0482] The "means for adjusting and sending a warning to a user terminal" is a mechanism for adjusting the content and display method of a warning message based on the calculated reliability score and the recognized emotional state of the user, and sending it to the user's terminal.

[0483] The "means for displaying the adjusted warning message on the user terminal" is a mechanism for displaying the adjusted warning message sent from the server on the user terminal.

[0484] "Means for receiving video or image data" refers to a mechanism for receiving video or image data uploaded by a user.

[0485] "Means for identifying traces of unauthorized editing" refers to a mechanism for analyzing received video or image data and checking for traces of editing or manipulation.

[0486] "Means for collecting site information" refers to a mechanism for collecting information on the social media sites that users access.

[0487] "Means for determining attempts at unauthorized data collection or tracking" refers to a mechanism for analyzing collected site information to determine whether the site is attempting unauthorized data collection or tracking of users.

[0488] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[0489] Evaluating the credibility of news articles and using sentiment engines

[0490] When a user browses a news article online, the user's device sends the URL of the news article to a server. The server retrieves the news article's text and metadata from the URL. A generative AI deployed on the server analyzes the news article and calculates a credibility score based on the text's content, the credibility of the source, and past history. Specifically, a generative AI model (such as GPT-4 for natural language processing) is used to analyze the text and assign a credibility score. An emotion engine (such as IBM Watson's Tone Analyzer) then recognizes the user's emotional state. The calculated credibility score and analysis results are sent to the user's device, and the display of the warning is adjusted based on the user's emotional state detected by the emotion engine. For example, when a warning message stating "This news article is unreliable" is displayed to a user, if the user is angry, the warning will be displayed more gently, whereas if the user is calm, a more direct warning will be displayed.

[0491] Example prompt sentence:

[0492] "Please parse the following news article URL and calculate a credibility score: [news article URL]"

[0493] Detecting edited video and image data and using emotion engines

[0494] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generated AI checks the video or image for signs of editing and identifies any unauthorized changes. For example, it uses Adobe Photoshop's image analysis function to detect editing. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. For example, the warning "This video may be faked" is adjusted according to the user's state.

[0495] Example prompt sentence:

[0496] "Analyze this video file to check for any signs of unauthorized editing: [video file path]"

[0497] Social media monitoring and sentiment engine utilization

[0498] When a user browses social media sites in a browser, their online activity is monitored using a browser extension. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. For example, this uses the monitoring function of Ghostery. The results of this determination are sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message urging them to stay calm is displayed. For example, the warning "This site is engaging in unauthorized data collection" is adjusted according to the user's state.

[0499] Example prompt sentence:

[0500] "Check if this social media site is illegally collecting data: [site URL]"

[0501] The goal of this system is to evaluate the reliability of news articles, video and image data, and social media information, helping users easily determine the accuracy of the information, while also taking into account the user's emotional state to provide more effective warnings.

[0502] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0503] Evaluating the credibility of news articles and using sentiment engines

[0504] Step 1:

[0505] A user clicks on a news article URL.

[0506] (Input) The URL the user clicks on.

[0507] (Output) URL information clicked on the user's device.

[0508] Users click on the URL of the news article they want to view in their web browser.

[0509] Step 2:

[0510] The device sends the URL information to the server.

[0511] (Input) The URL of the news article that was clicked.

[0512] (Output) The URL information sent to the server.

[0513] The user's terminal sends the URL of the clicked news article to the server as an HTTP request.

[0514] Step 3:

[0515] The server retrieves the news article text and metadata.

[0516] (Input) The URL of the received news article.

[0517] (Output) The news article text and metadata.

[0518] The server uses the received URL to obtain the text of the news article and metadata (publication date, author name, etc.) from the news site using scraping or an API.

[0519] Step 4:

[0520] The server instructs the generated AI to analyze.

[0521] (Input) News article text and metadata.

[0522] (Output) Analysis instructions to the generating AI.

[0523] The server sends the text and metadata of the retrieved news article as input to the generation AI, instructing it to analyze it.

[0524] Step 5:

[0525] The generative AI calculates a reliability score.

[0526] (Input) News article text and metadata.

[0527] (Output) The confidence score.

[0528] The generative AI (e.g., GPT-4) analyzes the text and metadata of the received news article and calculates a credibility score based on the content of the text, the reliability of the source, past history, etc.

[0529] Step 6:

[0530] An emotion engine recognizes the user's emotional state.

[0531] (Input) User behavior data (viewing time, scrolling speed, etc.).

[0532] (Output) The user's emotional state.

[0533] An emotion engine (e.g., IBM Watson Tone Analyzer) analyzes the user's behavioral data and recognizes the user's current emotional state (excited, angry, calm, etc.).

[0534] Step 7:

[0535] The server sends the analysis results and warning messages to the terminal.

[0536] (Input) Confidence score and user's emotional state.

[0537] (Output) The warning message.

[0538] The server adjusts the content of the warning message based on the reliability score calculated by the generation AI and the user's emotional state recognized by the emotion engine, and sends it to the user's device.

[0539] Step 8:

[0540] The terminal displays a warning message to the user.

[0541] (Input) The warning message received from the server.

[0542] (Output) The warning message to be displayed.

[0543] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the news article.

[0544] Detecting edited video and image data and using emotion engines

[0545] Step 1:

[0546] A user uploads video or image data to the system.

[0547] (Input) Video or image file.

[0548] (Output) The data uploaded to the system.

[0549] Users upload video or image files to the system via the Internet.

[0550] Step 2:

[0551] The device sends the data to the server.

[0552] (Input) Uploaded video or image data.

[0553] (Output) The data sent to the server.

[0554] The user's terminal transmits the uploaded video or image data to the server.

[0555] Step 3:

[0556] The server instructs the generated AI to analyze.

[0557] (Input) Received video or image data.

[0558] (Output) Analysis instructions to the generating AI.

[0559] The server sends the received data to the generation AI and instructs it to analyze it.

[0560] Step 4:

[0561] Generative AI checks for signs of editing in videos and images.

[0562] (Input) Video or image data.

[0563] (Output) Analysis results for edit traces.

[0564] The generating AI analyzes the received video and image data and checks for signs of editing or unauthorized processing.

[0565] Step 5:

[0566] An emotion engine recognizes the user's emotional state.

[0567] (Input) User behavior data (viewing time, reaction speed, etc.).

[0568] (Output) The user's emotional state.

[0569] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[0570] Step 6:

[0571] The server sends the analysis results and warning messages to the terminal.

[0572] (Input) Analysis results of editing traces and the user's emotional state.

[0573] (Output) The warning message.

[0574] The server adjusts the content of the warning message based on the analysis results and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[0575] Step 7:

[0576] The terminal displays a warning message to the user.

[0577] (Input) The warning message received from the server.

[0578] (Output) The warning message to be displayed.

[0579] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the video or image.

[0580] Social media monitoring and sentiment engine utilization

[0581] Step 1:

[0582] A user browses a social media site.

[0583] (Input) Social media site URL.

[0584] (Output) Site browsing information on the user's device.

[0585] Users use a web browser to browse social media sites.

[0586] Step 2:

[0587] The site information that the terminal is accessing is sent to the server.

[0588] (Input) Information about the site you are viewing.

[0589] (Output) Site information sent to the server.

[0590] The browser extension collects information about the social media sites you visit and sends it to a server.

[0591] Step 3:

[0592] The server analyzes the site information.

[0593] (Input) Received site information.

[0594] (Output) Site analysis results.

[0595] The server analyzes the received information from the social media site and evaluates its content.

[0596] Step 4:

[0597] The server determines whether or not unauthorized data collection has occurred.

[0598] (Input) Site analysis results.

[0599] (Output) The result of the invalid data collection determination.

[0600] The server uses the site information to determine whether unauthorized data collection or tracking is occurring.

[0601] Step 5:

[0602] An emotion engine recognizes the user's emotional state.

[0603] (Input) User behavior data (viewing time, reaction speed, etc.).

[0604] (Output) The user's emotional state.

[0605] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[0606] Step 6:

[0607] The server sends the judgment result and a warning message to the terminal.

[0608] (Input) The result of the fraudulent data collection and the user's emotional state.

[0609] (Output) The warning message.

[0610] The server adjusts the content of the warning message based on the judgment result and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[0611] Step 7:

[0612] The terminal displays a warning message to the user.

[0613] (Input) The warning message received from the server.

[0614] (Output) The warning message to be displayed.

[0615] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning about the trustworthiness of the social media site.

[0616] (Application example 2)

[0617] Next, a description will be given of Application Example 2. 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."

[0618] In today's Internet environment, there is a lot of false information and incorrect data, making it difficult for users to determine whether or not to trust this information. Displaying appropriate warning messages based on the user's emotional state is also problematic. Conventional systems display uniform warning messages that ignore the user's emotional state, which can prevent users from receiving warnings appropriately.

[0619] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring the text and metadata of news articles, means for analyzing the acquired text and metadata and calculating a reliability score of the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for recognizing the user's emotional state and adjusting the display of the warning message based on the emotional state, and means for analyzing the URL and metadata of video content, calculating a reliability score, and displaying a warning message. This allows users to quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into consideration, providing more effective and responsive warnings.

[0620] A "network" is an infrastructure that allows multiple computers and devices to communicate with each other.

[0621] "Data" is a coded set of letters and numbers that represent specific information.

[0622] A "news article" is a piece of writing that reports social, economic, political, or cultural events or information.

[0623] "Body" is the text portion that constitutes the main content of a news article.

[0624] "Metadata" is data that describes information and attributes about data.

[0625] A "trustworthiness score" is an evaluation index that quantifies the accuracy and reliability of information.

[0626] A "user terminal" is a device used by a user to obtain and operate information.

[0627] A "warning" is a message intended to inform the user of a potential risk or problem.

[0628] "Emotional state" refers to the user's current psychological response or emotional state.

[0629] A "warning message" is textual or graphical information displayed to inform the user of a warning.

[0630] "Video content" refers to media data that is viewed as moving images.

[0631] A "URL" is an address that identifies the location of a web page or other Internet resource.

[0632] "Analysis" is the process of examining data in detail to reveal its structure and meaning.

[0633] "Unauthorized editing" is any deliberate alteration made to a footage or image that is false or misleading.

[0634] "Online activities" refer to the various operations and actions that users perform using the Internet.

[0635] "Data collection" is the act of collecting and storing information for a specific purpose.

[0636] "Tracking" refers to any technology or process that follows or records a user's online actions or movements.

[0637] A "generative AI model" is a model that is trained to solve a specific problem using artificial intelligence techniques.

[0638] A "prompt" is a text instruction entered into a generative AI model.

[0639] This invention relates to a system that evaluates the reliability of information in content distribution services and displays appropriate warnings according to the user's emotional state. This system provides users with a safe and reliable information environment by monitoring news articles, video content, video and image data, and social media sites.

[0640] Key Components

[0641] The main components of this system are as follows:

[0642] 1. Data collection method: A module for collecting URLs and metadata of news articles and video content. This includes a mechanism for sending data from the user's device to the server.

[0643] 2. Analysis: Analyzes the text and metadata of incoming news articles and video content and calculates a credibility score for the information, using a generative AI model.

[0644] 3. Emotion Recognition Means: A module that recognizes the user's current emotional state. This includes emotion recognition software.

[0645] 4. Warning display means: A module for displaying appropriate warning messages to users based on their confidence scores and emotional states.

[0646] Hardware and software used

[0647] Hardware: User devices (smartphones, tablets, PCs, smart glasses), servers.

[0648] software:

[0649] Emotion recognition software (e.g., Microsoft Azure Emotion API)

[0650] Video analysis software (e.g., Google Cloud Video Intelligence API)

[0651] Front-end application for displaying alerts (e.g. React Native)

[0652] Data processing and data calculation

[0653] The server retrieves the URLs and metadata of news articles and video content and analyzes them using a generative AI model. Specifically, it uses the Google Cloud Video Intelligence API to analyze the video content and calculate a reliability score. For emotion recognition, it uses the Microsoft Azure Emotion API to detect the user's emotional state. The calculated reliability score and analysis results are sent to the user's device, and the display of warning messages is adjusted based on the user's emotional state.

[0654] Specific examples

[0655] When a user watches a news video on a content delivery service, the video's URL and metadata are sent to a server. The server analyzes the video using the Google Cloud Video Intelligence API, and if the video is deemed to have a low reliability score, it uses emotion recognition software to assess the user's emotional state. For example, if the user is excited, the system displays a mild-toned warning message saying, "Please watch with caution. This video may contain unreliable information." If the user is calm, a more direct warning message is displayed.

[0656] Prompt Sentence Examples

[0657] "This video may be unreliable. Please watch with caution. This video may contain unreliable information."

[0658] This allows users to quickly evaluate the accuracy of information and use the internet safely. Furthermore, the emotion engine optimizes warning messages, allowing users to receive information appropriately.

[0659] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0660] Step 1:

[0661] When a user watches video content on a content distribution service, the URL and metadata of that video are sent from the user's device to the server. Specifically, when the user presses the watch button, the URL and metadata are collected and sent to the server via the network. This allows the server to obtain the video information to analyze.

[0662] Step 2:

[0663] The server passes the received video URL and metadata to the Google Cloud Video Intelligence API, which analyzes the video content. The analysis calculates a credibility score based on the video's content, source credibility, past history, etc. The input for this step is the video URL and metadata, and the output is a credibility score.

[0664] Step 3:

[0665] The server sends the calculated reliability score to the user's device. At the same time, emotion recognition software (Microsoft Azure Emotion API) is launched on the user's device and recognizes the user's current emotional state through the camera. The inputs of this step are the reliability score and the user's camera image, and the output is the user's emotional state data.

[0666] Step 4:

[0667] The emotion recognition software analyzes the user's emotional state and sends the results to the server. Specifically, it identifies emotions from the user's facial expressions and vocal tone and returns the analysis results to the server. The input for this step is the camera footage, and the output is the recognized emotional state data.

[0668] Step 5:

[0669] The server adjusts the content and tone of the warning message based on the reliability score and the user's emotional state. For example, if the reliability score is low and the user is in an excited state, the warning message will read, "Please watch with caution. This video may contain unreliable information." The inputs of this step are the reliability score and the user's emotional state, and the output is the adjusted warning message.

[0670] Step 6:

[0671] Finally, the adjusted warning message is sent to the user terminal and displayed on the video content. The user can check the warning message while watching the video. The input of this step is the adjusted warning message, and the output is the display of the warning message to the user.

[0672] 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the 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.

[0673] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0674] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0675] [Second embodiment]

[0676] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0677] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0678] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0679] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0680] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0681] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0682] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0683] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0684] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[0685] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0686] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0687] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0688] ---

[0689] The present invention relates to a system for ensuring the accuracy of information on the Internet, and can be implemented as follows: The system of the present invention aims to evaluate and analyze news articles, video and image data, and information through social media to ensure their reliability and accuracy.

[0690] News article credibility assessment

[0691] When a user views a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the news article text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the content of the text, the reliability of the source, past history, etc. The calculated reliability score and analysis results are then sent to the user's device, and a warning message such as "This news article has low reliability" is displayed on the user's device.

[0692] Specific examples

[0693] When a user clicks on a news article shared on a social networking site, the URL is sent to the server, which analyzes the news article and displays a warning message on the user's device if the article has a low reliability score. In this case, the user can confirm the reliability of the news article.

[0694] Detecting edited video and image data

[0695] When a user browses video or image files online, they can upload them to the system. The user's device sends the video or image data to the server, which analyzes it. The server's AI then checks the video or image for signs of editing and identifies unnatural changes, such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message is displayed, such as "This video may be faked."

[0696] Specific examples

[0697] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If any unauthorized editing is detected, a warning is displayed on the user's device.

[0698] Social Media Monitoring

[0699] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed in real time.

[0700] Specific examples

[0701] As users browse social media sites, the browser extension monitors the site's scripts and requests, and if the server determines that the site is engaging in unauthorized data collection, a warning message is displayed on the user's device.

[0702] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection, allowing them to enjoy a safer and more reliable information environment.

[0703] ---

[0704] This concludes the description of the present invention, which provides concrete steps to improve the reliability of information in news articles, video and image data, and social media monitoring.

[0705] The processing flow will be explained below.

[0706] News article credibility assessment

[0707] Step 1:

[0708] A user clicks on a link to a news article. The user's device sends this link information to the server.

[0709] Step 2:

[0710] The server retrieves the HTML data of the news article from the received URL.

[0711] Step 3:

[0712] The server extracts the news article text and metadata from the HTML data it has obtained.

[0713] Step 4:

[0714] The generation AI placed on the server analyzes the text of the news article using natural language processing (NLP) technology and evaluates the article's content, source reliability, past history, etc.

[0715] Step 5:

[0716] The server calculates a reliability score based on the analysis results.

[0717] Step 6:

[0718] The server sends the calculated reliability score and a summary of the analysis results to the user's device.

[0719] Step 7:

[0720] The analysis results received by the user's device are displayed, and if necessary, a warning message such as "This news article is unreliable" is displayed.

[0721] Detecting edited video and image data

[0722] Step 1:

[0723] A user uploads video or image data to the system.

[0724] Step 2:

[0725] The user's terminal transmits the uploaded file data to the server.

[0726] Step 3:

[0727] The server receives the file data.

[0728] Step 4:

[0729] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[0730] Step 5:

[0731] The server sends the analysis results to the user's device.

[0732] Step 6:

[0733] The analysis results received by the user's device are displayed, along with a warning message such as "This video may be faked."

[0734] Social Media Monitoring

[0735] Step 1:

[0736] A user browses to a social media site.

[0737] Step 2:

[0738] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[0739] Step 3:

[0740] The user's terminal transmits the collected site information to the server.

[0741] Step 4:

[0742] The server analyzes the received site information.

[0743] Step 5:

[0744] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[0745] Step 6:

[0746] The server sends the analysis results to the user's device.

[0747] Step 7:

[0748] The analysis results received by the user's device are displayed, and a warning message such as "This site is collecting data illegally" is displayed in real time.

[0749] ---

[0750] These are the specific processing steps for assessing the credibility of news articles, detecting edited video and image data, and monitoring social media. By following these steps, the system can efficiently evaluate information and provide necessary warnings to users.

[0751] Example 1

[0752] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0753] The problem is that it is difficult for users to easily determine the authenticity of news articles, video and image data, and information on social media, which is flooding the internet. This can lead to the spread of false information and the violation of privacy through the unauthorized collection of data.

[0754] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0755] In this invention, the server includes: means for transmitting a specific URL of a news article from a user's device to the server; means for receiving the transmitted URL at the server and acquiring the text and metadata of the news article; means for analyzing the acquired text and metadata using a generative AI model located on the server to calculate a reliability score of the information; means for transmitting the calculated reliability score and the analysis result to the user's device and displaying them as a warning message; means for receiving video or image data uploaded from the user's device and transmitting the received video or image data to the server to detect editing and unauthorized manipulation; means for analyzing the transmitted data using a generative AI model located on the server to identify traces of unauthorized editing; means for transmitting the analysis result to the user's device and displaying a warning message; means for monitoring the user's online activity using a browser extension and collecting information about the sites accessed; means for transmitting the collected site information to the server, analyzing it using the generative AI model located on the server, and determining attempts of unauthorized data collection or tracking; and means for transmitting the determination result to the user's device and displaying it as a warning message in real time. This allows users to quickly and easily evaluate the reliability of news articles, video and image data, and information on social media and protect themselves from unauthorized data collection.

[0756] A "news article" is an article or news report published on the Internet, including the text and metadata such as the title, author, and publication date.

[0757] "URL" is an abbreviation for Uniform Resource Locator, and is a character string that indicates the address of a specific resource on the Internet.

[0758] A "generative AI model" is an artificial intelligence model used to perform natural language processing and data analysis, such as a machine learning model that generates sentences and performs data analysis.

[0759] The "trustworthiness score" is a numerical representation of the reliability and accuracy of news articles, video and image data, social media information, etc., and is an indicator of their reliability.

[0760] "Analysis" is the process of using generative AI models to examine the content of data in detail and evaluate its content, the reliability of its sources, and evidence of editing.

[0761] A "warning message" is a message that displays important information or a warning to the user, and is particularly a warning about information that is unreliable or inaccurate.

[0762] "Video data" refers to digital data stored in a moving image format, including video clips and movies.

[0763] "Image data" refers to digital data stored in still image format, including photographs and drawings.

[0764] "Editing traces" are unnatural changes or inconsistencies that serve as evidence that video or image data has been edited or manipulated.

[0765] A "browser extension" is additional software that extends the functionality of a web browser and has the ability to monitor a user's online activities and collect specific site information.

[0766] "Unauthorized data collection" refers to the act of illegally collecting personal data or usage information without the user's consent.

[0767] "Tracking" refers to the act of tracking a user's website browsing behavior and collecting and analyzing that data.

[0768] This invention relates to a system that monitors news articles, video and image data, and social media to evaluate the accuracy and reliability of the information. This system uses a generative AI model to evaluate the reliability of news articles, detect editing traces in video and image data, and monitor social media.

[0769] News article credibility assessment

[0770] When a user browses a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the text and metadata of the news article based on the received URL. A generative AI model (e.g., GPT-4) deployed on the server analyzes the text and metadata of the retrieved news article and calculates a reliability score. The analysis result and reliability score are then sent back to the user's device and displayed as a warning message on the user interface.

[0771] Specific examples

[0772] When a user clicks on a news article shared on a social networking site, the URL is sent to a server. The server analyzes the news article and calculates a reliability score using a generative AI model. If the reliability score is low, a warning message stating "This news article is unreliable" is displayed on the user's device.

[0773] Prompt Sentence Examples

[0774] "Analyze the text of news articles and calculate a credibility score."

[0775] Detecting edited video and image data

[0776] When a user views video or image files on the internet, they upload these files to the system. The user's device sends the video or image data to a server, which analyzes the data. A generative AI model (e.g., DALL-E) located on the server checks the video or image for signs of editing and identifies unnatural changes such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message such as "This video may be faked" is displayed.

[0777] Specific examples

[0778] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If unauthorized editing is detected, a warning message is displayed on the user's device stating, "This video may have been forged."

[0779] Prompt Sentence Examples

[0780] "Please detect any editing signs in this image and assess its authenticity."

[0781] Social Media Monitoring

[0782] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether or not unauthorized data collection or tracking is taking place. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is unauthorized data collection" is displayed in real time.

[0783] Specific examples

[0784] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's device will display a warning message stating, "This site is engaging in unauthorized data collection."

[0785] Prompt Sentence Examples

[0786] "Check if this site is engaging in unauthorized data collection or tracking."

[0787] Through this system, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling users to enjoy a safer and more reliable information environment.

[0788] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0789] News article credibility assessment

[0790] Step 1:

[0791] When a user views a news article, they click on the news article's URL.

[0792] Specific actions

[0793] When a user clicks on a news article link in an internet browser, JavaScript runs and retrieves the URL of the news article.

[0794] input

[0795] The URL of the news article the user clicked on.

[0796] output

[0797] The URL of the news article is retrieved on the user's device.

[0798] Step 2:

[0799] The device sends the URL of the news article to the server.

[0800] Specific actions

[0801] The JavaScript code sends the obtained URL to the backend server as a POST request.

[0802] input

[0803] The URL of the news article.

[0804] output

[0805] The URL of the news article is sent to the server.

[0806] Step 3:

[0807] The server retrieves the text and metadata of the news article based on the received URL.

[0808] Specific actions

[0809] The server uses Python's requests library to send requests to news sites, and extracts the text and metadata from the HTML using BeautifulSoup or similar.

[0810] input

[0811] The URL of the news article.

[0812] output

[0813] News article text and metadata.

[0814] Step 4:

[0815] The server sends the acquired news article text and metadata to the generative AI model, which generates an analysis prompt.

[0816] Specific actions

[0817] The server inputs the text and metadata of the news article into a generative AI (e.g., GPT-4) and generates a prompt saying, "Please rate the credibility of this article."

[0818] input

[0819] News article text and metadata.

[0820] output

[0821] The generated analysis prompt.

[0822] Step 5:

[0823] A generative AI model analyzes news articles and calculates a credibility score.

[0824] Specific actions

[0825] The generative AI model analyzes the text and metadata of the input news article and calculates a credibility score.

[0826] input

[0827] Parse prompts, news article body text and metadata.

[0828] output

[0829] Reliability scores and analysis results.

[0830] Step 6:

[0831] The server transmits the calculated reliability score and the analysis result to the user terminal.

[0832] Specific actions

[0833] The server returns the reliability score and analysis results in JSON format to the user device via the REST API set at the endpoint.

[0834] input

[0835] Reliability scores and analysis results.

[0836] output

[0837] The reliability score and analysis results sent to the user's device.

[0838] Step 7:

[0839] Based on the information received by the terminal, a warning message is displayed on the user interface.

[0840] Specific actions

[0841] The user interface (HTML, CSS, JavaScript) receives the reliability score and analysis results from the server and displays a warning message such as "This news article is not reliable" as a popup or banner.

[0842] input

[0843] Reliability scores and analysis results.

[0844] output

[0845] The warning message displayed in the user interface.

[0846] Detecting edited video and image data

[0847] Step 1:

[0848] A user uploads a video or image file onto the Internet.

[0849] Specific actions

[0850] The user interface displays a file selection dialog and the user selects a file.

[0851] input

[0852] Video or image files selected by the user.

[0853] output

[0854] The selected file will be uploaded to the system.

[0855] Step 2:

[0856] The terminal transmits the selected video or image data to the server.

[0857] Specific actions

[0858] Use the JavaScript FormData object to send the selected file to the server via a POST request.

[0859] input

[0860] Video or image files.

[0861] output

[0862] Video or image files sent to the server.

[0863] Step 3:

[0864] The server sends the received video or image data to the generative AI model, which generates an analysis prompt.

[0865] Specific actions

[0866] The server inputs the selected file data into a generation AI (e.g., DALL-E) and generates a prompt saying, "Please detect any editing traces in this image."

[0867] input

[0868] Uploaded video or image files.

[0869] output

[0870] The generated analysis prompt.

[0871] Step 4:

[0872] The generative AI model checks the video or image data for signs of editing and sends the results back to the server.

[0873] Specific actions

[0874] The AI ​​detects unnatural changes, color inconsistencies, and shadow discontinuities in the image and sends the analysis results back to the server.

[0875] input

[0876] Video or image file, analysis prompt.

[0877] output

[0878] Analysis results including traces of editing.

[0879] Step 5:

[0880] The server transmits the analysis results to the user terminal.

[0881] Specific actions

[0882] The server returns the analysis results in JSON format to the user device via the REST API set at the endpoint.

[0883] input

[0884] Analysis results including traces of editing.

[0885] output

[0886] Analysis results sent to the user's device.

[0887] Step 6:

[0888] Based on the analysis results, the device displays a warning message on the user interface.

[0889] Specific actions

[0890] The user interface receives the analysis results and displays a warning message such as "This footage may be faked" as a pop-up or banner.

[0891] input

[0892] Analysis results including traces of editing.

[0893] output

[0894] The warning message displayed in the user interface.

[0895] Social Media Monitoring

[0896] Step 1:

[0897] The browser extension monitors your online activity as you browse social media sites.

[0898] Specific actions

[0899] The browser extension collects the URLs of the sites you visit and the scripts those sites run.

[0900] input

[0901] Information about the social media sites you visit.

[0902] output

[0903] Site Information Collected.

[0904] Step 2:

[0905] The terminal sends the collected site information to the server.

[0906] Specific actions

[0907] The browser extension sends the collected data in JSON format to the server via a POST request.

[0908] input

[0909] Site Information Collected.

[0910] output

[0911] Site information sent to the server.

[0912] Step 3:

[0913] The server analyzes the site information it receives to determine whether unauthorized data collection or tracking is occurring.

[0914] Specific actions

[0915] The server analyzes the collected data against industry-standard security rules (for example, the OWASP threat list).

[0916] input

[0917] Received site information.

[0918] output

[0919] Verification of fraudulent data collection or tracking.

[0920] Step 4:

[0921] The server transmits the determination result to the user terminal.

[0922] Specific actions

[0923] The server returns the result of the judgment in JSON format to the user device via the REST API set at the endpoint.

[0924] input

[0925] Verification of fraudulent data collection or tracking.

[0926] output

[0927] The judgment result sent to the user terminal.

[0928] Step 5:

[0929] Based on the judgment result, the terminal displays a warning message on the user interface.

[0930] Specific actions

[0931] The user interface receives the judgment result and displays a warning message such as "This site is collecting data illegally" as a pop-up or banner.

[0932] input

[0933] Judgment result.

[0934] output

[0935] The warning message displayed in the user interface.

[0936] (Application example 1)

[0937] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0938] The decline in the reliability of information on the Internet and the resulting spread of misinformation and fake news have become a social problem. Furthermore, the increasing number of edited videos and images, as well as the increasing number of illegal data collection and tracking practices on social media, are making it difficult for users to determine the authenticity of information. These problems must be resolved quickly and accurately, and users must be provided with reliable information.

[0939] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0940] In this invention, the server includes means for collecting data via a network and acquiring news article text and metadata, means for analyzing the acquired text and metadata and calculating a reliability score for the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for receiving and analyzing video or image data and detecting editing and unauthorized manipulation, and means for monitoring and determining unauthorized data collection and tracking on social media sites and displaying a warning to the user. This allows users to quickly evaluate the reliability of information on the Internet and be protected from unauthorized data collection.

[0941] A "network" is a collection of computers and communication devices that allow information and data to be sent and received.

[0942] "Data" refers to the information being analyzed, such as news articles, videos, images, and social media information.

[0943] A "news article" is written information such as a news report posted on the Internet.

[0944] "Body" refers to the main content of a news article.

[0945] "Metadata" is supplemental information and attribute data associated with a news article.

[0946] A "trust score" is a numerical assessment of the trustworthiness of a particular article or piece of information.

[0947] A "warning" is a message that is displayed to the user to call attention to the situation.

[0948] "Video" is data that contains moving visual information.

[0949] An "image" is data containing still visual information.

[0950] "Editing evidence" refers to evidence of changes that indicate an image or video has been modified.

[0951] "Social media" is an online platform for sharing information between users.

[0952] "Unauthorized data collection" refers to the illegal act of collecting and using user information without permission.

[0953] "Tracking" is the act of following a user's online activities.

[0954] A "user terminal" is a device used to view information and receive alerts.

[0955] A "generative AI model" is a machine learning model for performing natural language processing and data analysis.

[0956] A "prompt sentence" is a sentence used to input instructions or questions into a generative AI model.

[0957] This invention provides a system for evaluating the reliability of information on the Internet. The system analyzes news articles, video and image data, and information on social media, and conveys the reliability to users.

[0958] Program Generation

[0959] The following describes an outline of a program and its processing for realizing the present invention.

[0960] Hardware / Software used

[0961] Hardware: User's smartphone, server

[0962] Software: Mobile apps, browser extensions, generative AI models (e.g., GPT-4), data analysis tools

[0963] News article credibility assessment

[0964] When a user views a news article on their smartphone, the URL of that news article is automatically sent to a server. A generative AI model (e.g., GPT-4) on the server retrieves and analyzes the news article text and metadata from the URL. Based on the analysis results, a reliability score is calculated and sent to the user's smartphone. If the reliability score is low, the user's smartphone displays a warning message such as "This news article is unreliable."

[0965] Examples:

[0966] When a user clicks on a news article shared on a social networking site, the URL is sent to a server, which analyzes the news article and, if the reliability score is low, displays a warning message on the user's smartphone saying, "This news article is unreliable."

[0967] Prompt for the generative AI model:

[0968] News article URL: [example.com / news123]

[0969] Rate the article's credibility and calculate a credibility score, and explain why it contains exaggerated or unreliable content.

[0970] Detecting traces of editing of video and image data

[0971] When a user views a video or image file on their smartphone, the file is automatically uploaded to a server. A generative AI model on the server analyzes the video or image and detects any signs of editing. If there is a problem with its authenticity, that information is sent to the user's smartphone, warning them that "this video may be faked."

[0972] Examples:

[0973] When a user downloads a video shared on a social networking site and uploads it to the system, the server analyzes the video, and if any fraud is detected, a warning message will appear on the smartphone saying, "This video may be faked."

[0974] Social Media Monitoring

[0975] When a user visits a social media site in their browser, the browser extension monitors the user's online activity and sends the site information to a server. An analysis tool on the server identifies any unauthorized data collection or tracking attempts and sends the results to the user's smartphone. If any unauthorized activity is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed.

[0976] Examples:

[0977] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's smartphone will display a warning message stating, "This site is engaging in unauthorized data collection."

[0978] In this way, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling them to enjoy a safer and more reliable information environment.

[0979] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0980] Program flow:

[0981] News article credibility assessment

[0982] Step 1:

[0983] The user terminal inputs the URL of the news article.

[0984] The entered URL is sent from the user's smartphone to the server.

[0985] Step 2:

[0986] The server retrieves the text and metadata of the news article based on the received URL.

[0987] The server uses a web scraping tool to retrieve data from the URL.

[0988] Step 3:

[0989] The server parses the retrieved text and metadata.

[0990] A generative AI model (e.g., GPT-4) is used to calculate a confidence score for the text.

[0991] Step 4:

[0992] The server transmits the calculated reliability score to the user terminal.

[0993] The reliability score and analysis results are displayed on the user's smartphone.

[0994] Step 5:

[0995] If the user terminal has a low reliability score, a warning message is displayed.

[0996] Users may see a message that reads, "This news article is not reliable."

[0997] Detecting traces of editing of video and image data

[0998] Step 1:

[0999] The user uploads a video or image to the device.

[1000] The uploaded data is automatically sent to the server.

[1001] Step 2:

[1002] The server analyzes the received video or image data.

[1003] Generative AI models are used to detect edits (e.g., unnatural color tones and shadows).

[1004] Step 3:

[1005] The server transmits the analysis results to the user terminal.

[1006] A warning message appears on the user's smartphone saying, "This video may be faked."

[1007] Step 4:

[1008] The user terminal displays a warning message and presents the analysis results.

[1009] Users can see the specific locations and reasons for unauthorized edits.

[1010] Social Media Monitoring

[1011] Step 1:

[1012] A user browses a social media site.

[1013] Browser extensions monitor your online activity and collect site information.

[1014] Step 2:

[1015] The server analyzes the collected site information.

[1016] Generative AI models identify potential fraudulent data collection and tracking.

[1017] Step 3:

[1018] The server transmits the determination result to the user terminal.

[1019] A warning message appears on the user's smartphone saying, "This site is collecting data illegally."

[1020] Step 4:

[1021] The user terminal displays a warning message and provides details of the fraudulent activity.

[1022] Users will be given a detailed explanation of what fraudulent activity is taking place.

[1023] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1024] ---

[1025] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[1026] Evaluating the credibility of news articles and using sentiment engines

[1027] When a user views a news article online, the user's device sends the news article's URL to a server. The server retrieves the news article's text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the text content, source reliability, past history, etc. The user's emotion engine then recognizes the user's emotional state (excited, angry, calm, etc.). The calculated reliability score and analysis results are sent to the user's device, and the way the warning is displayed is adjusted based on the user's emotional state detected by the emotion engine. When a warning message such as "This news article is unreliable" is displayed, if the user is angry, the warning will be displayed more gently, and if the user is calm, a more direct warning will be displayed.

[1028] Specific examples

[1029] When a user clicks on a news article shared on a social networking site, the URL is sent to the server. The server analyzes the news article, and if the reliability score is low, a warning message is displayed on the user's device. At the same time, the emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a gentle warning if the user is excited.

[1030] Detecting edited video and image data and using emotion engines

[1031] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generation AI checks the video or image for signs of editing and identifies any unauthorized changes. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. The warning, "This video may be faked," is adjusted according to the user's state.

[1032] Specific examples

[1033] Users download videos posted on social media and upload them to the system. The server analyzes the video to determine whether there are any signs of unauthorized editing. If unauthorized editing is detected, the emotion engine recognizes the user's emotions and displays a gentle warning message if the user is feeling angry.

[1034] Social media monitoring and sentiment engine utilization

[1035] When a user visits a social media site in a browser, a browser extension is used to monitor the user's online activity. The user's device sends information about the site being accessed to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The result of the determination is sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message is displayed encouraging them to stay calm. The warning, "This site is engaging in unauthorized data collection," is adjusted according to the user's state.

[1036] Specific examples

[1037] As a user browses social media sites, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, an emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a more gentle message if the user is impatient.

[1038] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into account, providing more effective and responsive warnings. This allows users to enjoy a safer and more reliable information environment.

[1039] ---

[1040] This concludes the description of the embodiment of the present invention. This system can more effectively ensure the accuracy of information and provide appropriate warnings to users by evaluating the reliability of news articles, detecting edits in video and image data, and monitoring social media, as well as taking into account the emotional state of the user.

[1041] The processing flow will be explained below.

[1042] Evaluating the credibility of news articles and using sentiment engines

[1043] Step 1:

[1044] A user clicks on a news article link, and the user's device sends the news article URL to the server.

[1045] Step 2:

[1046] The server retrieves the HTML data of the news article from the received URL.

[1047] Step 3:

[1048] The server extracts the news article text and metadata from the HTML data it has obtained.

[1049] Step 4:

[1050] The generation AI placed on the server analyzes the text of news articles using natural language processing (NLP) technology and evaluates the content of the article, the reliability of the source, past history, etc.

[1051] Step 5:

[1052] The server calculates a reliability score based on the analysis results.

[1053] Step 6:

[1054] The server queries the user's emotion engine for the user's most recent emotional state.

[1055] Step 7:

[1056] The user's device activates an emotion engine to detect the user's emotional state (excitement, anger, calmness, etc.).

[1057] Step 8:

[1058] The emotion engine transmits the user's emotional state to the server.

[1059] Step 9:

[1060] The server adjusts how the warning message is displayed depending on the confidence score and the detected emotional state.

[1061] Step 10:

[1062] The server sends a reliability score and a warning message to the user's device.

[1063] Step 11:

[1064] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[1065] Detecting edited video and image data and using emotion engines

[1066] Step 1:

[1067] A user uploads video or image data to the system.

[1068] Step 2:

[1069] The user's terminal transmits the uploaded file data to the server.

[1070] Step 3:

[1071] The server receives the file data.

[1072] Step 4:

[1073] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[1074] Step 5:

[1075] The server creates a warning message based on the analysis results.

[1076] Step 6:

[1077] The server queries the user's emotion engine for the emotional state.

[1078] Step 7:

[1079] The user's terminal activates an emotion engine to detect the user's emotional state.

[1080] Step 8:

[1081] The emotion engine transmits the user's emotional state to the server.

[1082] Step 9:

[1083] The server tailors the warning message based on the analysis results and emotional state.

[1084] Step 10:

[1085] The server sends the analysis results and warning messages to the user's terminal.

[1086] Step 11:

[1087] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[1088] Social media monitoring and sentiment engine utilization

[1089] Step 1:

[1090] A user browses to a social media site.

[1091] Step 2:

[1092] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[1093] Step 3:

[1094] The user's terminal transmits the collected site information to the server.

[1095] Step 4:

[1096] The server analyzes the received site information.

[1097] Step 5:

[1098] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[1099] Step 6:

[1100] The server generates a verdict and a warning message.

[1101] Step 7:

[1102] The server queries the user's emotion engine for the emotional state.

[1103] Step 8:

[1104] The user's terminal activates an emotion engine to detect the user's emotional state.

[1105] Step 9:

[1106] The emotion engine transmits the user's emotional state to the server.

[1107] Step 10:

[1108] The server adjusts the warning message based on the result of the assessment and the emotional state.

[1109] Step 11:

[1110] The server sends the judgment result and a warning message to the user's terminal.

[1111] Step 12:

[1112] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[1113] ---

[1114] In this way, by taking into account the user's emotional state in addition to assessing the reliability of news articles, detecting traces of editing in video and image data, and monitoring social media, we can more effectively ensure the accuracy of information and provide appropriate warnings to users.

[1115] Example 2

[1116] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1117] The internet is flooded with information, often containing unreliable news articles and fraudulently manipulated video and images. Social media sites also sometimes conduct fraudulent data collection and tracking. This information not only causes confusion and anxiety among users, but can also lead to social unrest. Therefore, it is important to evaluate the reliability of this information and provide appropriate warnings to users. However, conventional systems often display uniform warnings without considering the user's emotional state, which can detract from the user experience.

[1118] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1119] In this invention, the server includes means for acquiring news article text and metadata, means for analyzing the acquired text and metadata and using a generative AI model to calculate a reliability score for the information, means for using an emotion engine to recognize the user's emotional state, means for adjusting a warning based on the calculated reliability score and the user's emotional state and sending the adjusted warning message to the user terminal, and means for displaying the adjusted warning message on the user terminal. This makes it possible to evaluate the reliability of news articles, video / image data, and social media site information, and to provide flexible warnings that take the user's emotional state into consideration.

[1120] "Means for obtaining news article text and metadata" refers to a system for collecting the text of news articles published on the Internet and related metadata such as publication date, author information, and URL.

[1121] "Means for using generative AI models" refers to a mechanism for using generative artificial intelligence models to analyze collected data and output the results.

[1122] An "emotion engine for recognizing the user's emotional state" is an engine that can identify and analyze the user's emotions, and is a mechanism for determining whether the user is excited, angry, calm, etc.

[1123] The "means for calculating the reliability score of information" is a system that analyzes the content and metadata of news articles, video and image data, and expresses the reliability of that information as a quantified score.

[1124] The "means for adjusting and sending a warning to a user terminal" is a mechanism for adjusting the content and display method of a warning message based on the calculated reliability score and the recognized emotional state of the user, and sending it to the user's terminal.

[1125] The "means for displaying the adjusted warning message on the user terminal" is a mechanism for displaying the adjusted warning message sent from the server on the user terminal.

[1126] "Means for receiving video or image data" refers to a mechanism for receiving video or image data uploaded by a user.

[1127] "Means for identifying traces of unauthorized editing" refers to a mechanism for analyzing received video or image data and checking for traces of editing or manipulation.

[1128] "Means for collecting site information" refers to a mechanism for collecting information on the social media sites that users access.

[1129] "Means for determining attempts at unauthorized data collection or tracking" refers to a mechanism for analyzing collected site information to determine whether the site is attempting unauthorized data collection or tracking of users.

[1130] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[1131] Evaluating the credibility of news articles and using sentiment engines

[1132] When a user browses a news article online, the user's device sends the URL of the news article to a server. The server retrieves the news article's text and metadata from the URL. A generative AI deployed on the server analyzes the news article and calculates a credibility score based on the text's content, the credibility of the source, and past history. Specifically, a generative AI model (such as GPT-4 for natural language processing) is used to analyze the text and assign a credibility score. An emotion engine (such as IBM Watson's Tone Analyzer) then recognizes the user's emotional state. The calculated credibility score and analysis results are sent to the user's device, and the display of the warning is adjusted based on the user's emotional state detected by the emotion engine. For example, when a warning message stating "This news article is unreliable" is displayed to a user, if the user is angry, the warning will be displayed more gently, whereas if the user is calm, a more direct warning will be displayed.

[1133] Example prompt sentence:

[1134] "Please parse the following news article URL and calculate a credibility score: [news article URL]"

[1135] Detecting edited video and image data and using emotion engines

[1136] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generated AI checks the video or image for signs of editing and identifies any unauthorized changes. For example, it uses Adobe Photoshop's image analysis function to detect editing. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. For example, the warning "This video may be faked" is adjusted according to the user's state.

[1137] Example prompt sentence:

[1138] "Analyze this video file to check for any signs of unauthorized editing: [video file path]"

[1139] Social media monitoring and sentiment engine utilization

[1140] When a user browses social media sites in a browser, their online activity is monitored using a browser extension. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. For example, this uses the monitoring function of Ghostery. The results of this determination are sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message urging them to stay calm is displayed. For example, the warning "This site is engaging in unauthorized data collection" is adjusted according to the user's state.

[1141] Example prompt sentence:

[1142] "Check if this social media site is illegally collecting data: [site URL]"

[1143] The goal of this system is to evaluate the reliability of news articles, video and image data, and social media information, helping users easily determine the accuracy of the information, while also taking into account the user's emotional state to provide more effective warnings.

[1144] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1145] Evaluating the credibility of news articles and using sentiment engines

[1146] Step 1:

[1147] A user clicks on a news article URL.

[1148] (Input) The URL the user clicks on.

[1149] (Output) URL information clicked on the user's device.

[1150] Users click on the URL of the news article they want to view in their web browser.

[1151] Step 2:

[1152] The device sends the URL information to the server.

[1153] (Input) The URL of the news article that was clicked.

[1154] (Output) The URL information sent to the server.

[1155] The user's terminal sends the URL of the clicked news article to the server as an HTTP request.

[1156] Step 3:

[1157] The server retrieves the news article text and metadata.

[1158] (Input) The URL of the received news article.

[1159] (Output) The news article text and metadata.

[1160] The server uses the received URL to obtain the text of the news article and metadata (publication date, author name, etc.) from the news site using scraping or an API.

[1161] Step 4:

[1162] The server instructs the generated AI to analyze.

[1163] (Input) News article text and metadata.

[1164] (Output) Analysis instructions to the generating AI.

[1165] The server sends the text and metadata of the retrieved news article as input to the generation AI, instructing it to analyze it.

[1166] Step 5:

[1167] The generative AI calculates a reliability score.

[1168] (Input) News article text and metadata.

[1169] (Output) The confidence score.

[1170] The generative AI (e.g., GPT-4) analyzes the text and metadata of the received news article and calculates a credibility score based on the content of the text, the reliability of the source, past history, etc.

[1171] Step 6:

[1172] An emotion engine recognizes the user's emotional state.

[1173] (Input) User behavior data (viewing time, scrolling speed, etc.).

[1174] (Output) The user's emotional state.

[1175] An emotion engine (e.g., IBM Watson Tone Analyzer) analyzes the user's behavioral data and recognizes the user's current emotional state (excited, angry, calm, etc.).

[1176] Step 7:

[1177] The server sends the analysis results and warning messages to the terminal.

[1178] (Input) Confidence score and user's emotional state.

[1179] (Output) The warning message.

[1180] The server adjusts the content of the warning message based on the reliability score calculated by the generation AI and the user's emotional state recognized by the emotion engine, and sends it to the user's device.

[1181] Step 8:

[1182] The terminal displays a warning message to the user.

[1183] (Input) The warning message received from the server.

[1184] (Output) The warning message to be displayed.

[1185] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the news article.

[1186] Detecting edited video and image data and using emotion engines

[1187] Step 1:

[1188] A user uploads video or image data to the system.

[1189] (Input) Video or image file.

[1190] (Output) The data uploaded to the system.

[1191] Users upload video or image files to the system via the Internet.

[1192] Step 2:

[1193] The device sends the data to the server.

[1194] (Input) Uploaded video or image data.

[1195] (Output) The data sent to the server.

[1196] The user's terminal transmits the uploaded video or image data to the server.

[1197] Step 3:

[1198] The server instructs the generated AI to analyze.

[1199] (Input) Received video or image data.

[1200] (Output) Analysis instructions to the generating AI.

[1201] The server sends the received data to the generation AI and instructs it to analyze it.

[1202] Step 4:

[1203] Generative AI checks for signs of editing in videos and images.

[1204] (Input) Video or image data.

[1205] (Output) Analysis results for edit traces.

[1206] The generating AI analyzes the received video and image data and checks for signs of editing or unauthorized processing.

[1207] Step 5:

[1208] An emotion engine recognizes the user's emotional state.

[1209] (Input) User behavior data (viewing time, reaction speed, etc.).

[1210] (Output) The user's emotional state.

[1211] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[1212] Step 6:

[1213] The server sends the analysis results and warning messages to the terminal.

[1214] (Input) Analysis results of editing traces and the user's emotional state.

[1215] (Output) The warning message.

[1216] The server adjusts the content of the warning message based on the analysis results and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[1217] Step 7:

[1218] The terminal displays a warning message to the user.

[1219] (Input) The warning message received from the server.

[1220] (Output) The warning message to be displayed.

[1221] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the video or image.

[1222] Social media monitoring and sentiment engine utilization

[1223] Step 1:

[1224] A user browses a social media site.

[1225] (Input) Social media site URL.

[1226] (Output) Site browsing information on the user's device.

[1227] Users use a web browser to browse social media sites.

[1228] Step 2:

[1229] The site information that the terminal is accessing is sent to the server.

[1230] (Input) Information about the site you are viewing.

[1231] (Output) Site information sent to the server.

[1232] The browser extension collects information about the social media sites you visit and sends it to a server.

[1233] Step 3:

[1234] The server analyzes the site information.

[1235] (Input) Received site information.

[1236] (Output) Site analysis results.

[1237] The server analyzes the received information from the social media site and evaluates its content.

[1238] Step 4:

[1239] The server determines whether or not unauthorized data collection has occurred.

[1240] (Input) Site analysis results.

[1241] (Output) The result of the invalid data collection determination.

[1242] The server uses the site information to determine whether unauthorized data collection or tracking is occurring.

[1243] Step 5:

[1244] An emotion engine recognizes the user's emotional state.

[1245] (Input) User behavior data (viewing time, reaction speed, etc.).

[1246] (Output) The user's emotional state.

[1247] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[1248] Step 6:

[1249] The server sends the judgment result and a warning message to the terminal.

[1250] (Input) The result of the fraudulent data collection and the user's emotional state.

[1251] (Output) The warning message.

[1252] The server adjusts the content of the warning message based on the judgment result and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[1253] Step 7:

[1254] The terminal displays a warning message to the user.

[1255] (Input) The warning message received from the server.

[1256] (Output) The warning message to be displayed.

[1257] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning about the trustworthiness of the social media site.

[1258] (Application example 2)

[1259] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1260] In today's Internet environment, there is a lot of false information and incorrect data, making it difficult for users to determine whether or not to trust this information. Displaying appropriate warning messages based on the user's emotional state is also problematic. Conventional systems display uniform warning messages that ignore the user's emotional state, which can prevent users from receiving warnings appropriately.

[1261] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring the text and metadata of news articles, means for analyzing the acquired text and metadata and calculating a reliability score of the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for recognizing the user's emotional state and adjusting the display of the warning message based on the emotional state, and means for analyzing the URL and metadata of video content, calculating a reliability score, and displaying a warning message. This allows users to quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into consideration, providing more effective and responsive warnings.

[1262] A "network" is an infrastructure that allows multiple computers and devices to communicate with each other.

[1263] "Data" is a coded set of letters and numbers that represent specific information.

[1264] A "news article" is a piece of writing that reports social, economic, political, or cultural events or information.

[1265] "Body" is the text portion that constitutes the main content of a news article.

[1266] "Metadata" is data that describes information and attributes about data.

[1267] A "trustworthiness score" is an evaluation index that quantifies the accuracy and reliability of information.

[1268] A "user terminal" is a device used by a user to obtain and operate information.

[1269] A "warning" is a message intended to inform the user of a potential risk or problem.

[1270] "Emotional state" refers to the user's current psychological response or emotional state.

[1271] A "warning message" is textual or graphical information displayed to inform the user of a warning.

[1272] "Video content" refers to media data that is viewed as moving images.

[1273] A "URL" is an address that identifies the location of a web page or other Internet resource.

[1274] "Analysis" is the process of examining data in detail to reveal its structure and meaning.

[1275] "Unauthorized editing" is any deliberate alteration made to a footage or image that is false or misleading.

[1276] "Online activities" refer to the various operations and actions that users perform using the Internet.

[1277] "Data collection" is the act of collecting and storing information for a specific purpose.

[1278] "Tracking" refers to any technology or process that follows or records a user's online actions or movements.

[1279] A "generative AI model" is a model that is trained to solve a specific problem using artificial intelligence techniques.

[1280] A "prompt" is a text instruction entered into a generative AI model.

[1281] This invention relates to a system that evaluates the reliability of information in content distribution services and displays appropriate warnings according to the user's emotional state. This system provides users with a safe and reliable information environment by monitoring news articles, video content, video and image data, and social media sites.

[1282] Key Components

[1283] The main components of this system are as follows:

[1284] 1. Data collection method: A module for collecting URLs and metadata of news articles and video content. This includes a mechanism for sending data from the user's device to the server.

[1285] 2. Analysis: Analyzes the text and metadata of incoming news articles and video content and calculates a credibility score for the information, using a generative AI model.

[1286] 3. Emotion Recognition Means: A module that recognizes the user's current emotional state. This includes emotion recognition software.

[1287] 4. Warning display means: A module for displaying appropriate warning messages to users based on their confidence scores and emotional states.

[1288] Hardware and software used

[1289] Hardware: User devices (smartphones, tablets, PCs, smart glasses), servers.

[1290] software:

[1291] Emotion recognition software (e.g., Microsoft Azure Emotion API)

[1292] Video analysis software (e.g., Google Cloud Video Intelligence API)

[1293] Front-end application for displaying alerts (e.g. React Native)

[1294] Data processing and data calculation

[1295] The server retrieves the URLs and metadata of news articles and video content and analyzes them using a generative AI model. Specifically, it uses the Google Cloud Video Intelligence API to analyze the video content and calculate a reliability score. For emotion recognition, it uses the Microsoft Azure Emotion API to detect the user's emotional state. The calculated reliability score and analysis results are sent to the user's device, and the display of warning messages is adjusted based on the user's emotional state.

[1296] Specific examples

[1297] When a user watches a news video on a content delivery service, the video's URL and metadata are sent to a server. The server analyzes the video using the Google Cloud Video Intelligence API, and if the video is deemed to have a low reliability score, it uses emotion recognition software to assess the user's emotional state. For example, if the user is excited, the system displays a mild-toned warning message saying, "Please watch with caution. This video may contain unreliable information." If the user is calm, a more direct warning message is displayed.

[1298] Prompt Sentence Examples

[1299] "This video may be unreliable. Please watch with caution. This video may contain unreliable information."

[1300] This allows users to quickly evaluate the accuracy of information and use the internet safely. Furthermore, the emotion engine optimizes warning messages, allowing users to receive information appropriately.

[1301] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1302] Step 1:

[1303] When a user watches video content on a content distribution service, the URL and metadata of that video are sent from the user's device to the server. Specifically, when the user presses the watch button, the URL and metadata are collected and sent to the server via the network. This allows the server to obtain the video information to analyze.

[1304] Step 2:

[1305] The server passes the received video URL and metadata to the Google Cloud Video Intelligence API, which analyzes the video content. The analysis calculates a credibility score based on the video's content, source credibility, past history, etc. The input for this step is the video URL and metadata, and the output is a credibility score.

[1306] Step 3:

[1307] The server sends the calculated reliability score to the user's device. At the same time, emotion recognition software (Microsoft Azure Emotion API) is launched on the user's device and recognizes the user's current emotional state through the camera. The inputs of this step are the reliability score and the user's camera image, and the output is the user's emotional state data.

[1308] Step 4:

[1309] The emotion recognition software analyzes the user's emotional state and sends the results to the server. Specifically, it identifies emotions from the user's facial expressions and vocal tone and returns the analysis results to the server. The input for this step is the camera footage, and the output is the recognized emotional state data.

[1310] Step 5:

[1311] The server adjusts the content and tone of the warning message based on the reliability score and the user's emotional state. For example, if the reliability score is low and the user is in an excited state, the warning message will read, "Please watch with caution. This video may contain unreliable information." The inputs of this step are the reliability score and the user's emotional state, and the output is the adjusted warning message.

[1312] Step 6:

[1313] Finally, the adjusted warning message is sent to the user terminal and displayed on the video content. The user can check the warning message while watching the video. The input of this step is the adjusted warning message, and the output is the display of the warning message to the user.

[1314] 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 a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1315] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1316] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1317] [Third embodiment]

[1318] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1319] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1320] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1321] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1322] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1323] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1324] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1325] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1326] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[1327] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1328] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1329] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[1330] ---

[1331] The present invention relates to a system for ensuring the accuracy of information on the Internet, and can be implemented as follows: The system of the present invention aims to evaluate and analyze news articles, video and image data, and information through social media to ensure their reliability and accuracy.

[1332] News article credibility assessment

[1333] When a user views a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the news article text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the content of the text, the reliability of the source, past history, etc. The calculated reliability score and analysis results are then sent to the user's device, and a warning message such as "This news article has low reliability" is displayed on the user's device.

[1334] Specific examples

[1335] When a user clicks on a news article shared on a social networking site, the URL is sent to the server, which analyzes the news article and displays a warning message on the user's device if the article has a low reliability score. In this case, the user can confirm the reliability of the news article.

[1336] Detecting edited video and image data

[1337] When a user browses video or image files online, they can upload them to the system. The user's device sends the video or image data to the server, which analyzes it. The server's AI then checks the video or image for signs of editing and identifies unnatural changes, such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message is displayed, such as "This video may be faked."

[1338] Specific examples

[1339] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If any unauthorized editing is detected, a warning is displayed on the user's device.

[1340] Social Media Monitoring

[1341] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed in real time.

[1342] Specific examples

[1343] As users browse social media sites, the browser extension monitors the site's scripts and requests, and if the server determines that the site is engaging in unauthorized data collection, a warning message is displayed on the user's device.

[1344] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection, allowing them to enjoy a safer and more reliable information environment.

[1345] ---

[1346] This concludes the description of the present invention, which provides concrete steps to improve the reliability of information in news articles, video and image data, and social media monitoring.

[1347] The processing flow will be explained below.

[1348] News article credibility assessment

[1349] Step 1:

[1350] A user clicks on a link to a news article. The user's device sends this link information to the server.

[1351] Step 2:

[1352] The server retrieves the HTML data of the news article from the received URL.

[1353] Step 3:

[1354] The server extracts the news article text and metadata from the HTML data it has obtained.

[1355] Step 4:

[1356] The generation AI placed on the server analyzes the text of the news article using natural language processing (NLP) technology and evaluates the article's content, source reliability, past history, etc.

[1357] Step 5:

[1358] The server calculates a reliability score based on the analysis results.

[1359] Step 6:

[1360] The server sends the calculated reliability score and a summary of the analysis results to the user's device.

[1361] Step 7:

[1362] The analysis results received by the user's device are displayed, and if necessary, a warning message such as "This news article is unreliable" is displayed.

[1363] Detecting edited video and image data

[1364] Step 1:

[1365] A user uploads video or image data to the system.

[1366] Step 2:

[1367] The user's terminal transmits the uploaded file data to the server.

[1368] Step 3:

[1369] The server receives the file data.

[1370] Step 4:

[1371] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[1372] Step 5:

[1373] The server sends the analysis results to the user's device.

[1374] Step 6:

[1375] The analysis results received by the user's device are displayed, along with a warning message such as "This video may be faked."

[1376] Social Media Monitoring

[1377] Step 1:

[1378] A user browses to a social media site.

[1379] Step 2:

[1380] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[1381] Step 3:

[1382] The user's terminal transmits the collected site information to the server.

[1383] Step 4:

[1384] The server analyzes the received site information.

[1385] Step 5:

[1386] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[1387] Step 6:

[1388] The server sends the analysis results to the user's device.

[1389] Step 7:

[1390] The analysis results received by the user's device are displayed, and a warning message such as "This site is collecting data illegally" is displayed in real time.

[1391] ---

[1392] These are the specific processing steps for assessing the credibility of news articles, detecting edited video and image data, and monitoring social media. By following these steps, the system can efficiently evaluate information and provide necessary warnings to users.

[1393] Example 1

[1394] Next, a description will be given of Example 1. 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."

[1395] The problem is that it is difficult for users to easily determine the authenticity of news articles, video and image data, and information on social media, which is flooding the internet. This can lead to the spread of false information and the violation of privacy through the unauthorized collection of data.

[1396] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1397] In this invention, the server includes: means for transmitting a specific URL of a news article from a user's device to the server; means for receiving the transmitted URL at the server and acquiring the text and metadata of the news article; means for analyzing the acquired text and metadata using a generative AI model located on the server to calculate a reliability score of the information; means for transmitting the calculated reliability score and the analysis result to the user's device and displaying them as a warning message; means for receiving video or image data uploaded from the user's device and transmitting the received video or image data to the server to detect editing and unauthorized manipulation; means for analyzing the transmitted data using a generative AI model located on the server to identify traces of unauthorized editing; means for transmitting the analysis result to the user's device and displaying a warning message; means for monitoring the user's online activity using a browser extension and collecting information about the sites accessed; means for transmitting the collected site information to the server, analyzing it using the generative AI model located on the server, and determining attempts of unauthorized data collection or tracking; and means for transmitting the determination result to the user's device and displaying it as a warning message in real time. This allows users to quickly and easily evaluate the reliability of news articles, video and image data, and information on social media and protect themselves from unauthorized data collection.

[1398] A "news article" is an article or news report published on the Internet, including the text and metadata such as the title, author, and publication date.

[1399] "URL" is an abbreviation for Uniform Resource Locator, and is a character string that indicates the address of a specific resource on the Internet.

[1400] A "generative AI model" is an artificial intelligence model used to perform natural language processing and data analysis, such as a machine learning model that generates sentences and performs data analysis.

[1401] The "trustworthiness score" is a numerical representation of the reliability and accuracy of news articles, video and image data, social media information, etc., and is an indicator of their reliability.

[1402] "Analysis" is the process of using generative AI models to examine the content of data in detail and evaluate its content, the reliability of its sources, and evidence of editing.

[1403] A "warning message" is a message that displays important information or a warning to the user, and is particularly a warning about information that is unreliable or inaccurate.

[1404] "Video data" refers to digital data stored in a moving image format, including video clips and movies.

[1405] "Image data" refers to digital data stored in still image format, including photographs and drawings.

[1406] "Editing traces" are unnatural changes or inconsistencies that serve as evidence that video or image data has been edited or manipulated.

[1407] A "browser extension" is additional software that extends the functionality of a web browser and has the ability to monitor a user's online activities and collect specific site information.

[1408] "Unauthorized data collection" refers to the act of illegally collecting personal data or usage information without the user's consent.

[1409] "Tracking" refers to the act of tracking a user's website browsing behavior and collecting and analyzing that data.

[1410] This invention relates to a system that monitors news articles, video and image data, and social media to evaluate the accuracy and reliability of the information. This system uses a generative AI model to evaluate the reliability of news articles, detect editing traces in video and image data, and monitor social media.

[1411] News article credibility assessment

[1412] When a user browses a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the text and metadata of the news article based on the received URL. A generative AI model (e.g., GPT-4) deployed on the server analyzes the text and metadata of the retrieved news article and calculates a reliability score. The analysis result and reliability score are then sent back to the user's device and displayed as a warning message on the user interface.

[1413] Specific examples

[1414] When a user clicks on a news article shared on a social networking site, the URL is sent to a server. The server analyzes the news article and calculates a reliability score using a generative AI model. If the reliability score is low, a warning message stating "This news article is unreliable" is displayed on the user's device.

[1415] Prompt Sentence Examples

[1416] "Analyze the text of news articles and calculate a credibility score."

[1417] Detecting edited video and image data

[1418] When a user views video or image files on the internet, they upload these files to the system. The user's device sends the video or image data to a server, which analyzes the data. A generative AI model (e.g., DALL-E) located on the server checks the video or image for signs of editing and identifies unnatural changes such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message such as "This video may be faked" is displayed.

[1419] Specific examples

[1420] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If unauthorized editing is detected, a warning message is displayed on the user's device stating, "This video may have been forged."

[1421] Prompt Sentence Examples

[1422] "Please detect any editing signs in this image and assess its authenticity."

[1423] Social Media Monitoring

[1424] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether or not unauthorized data collection or tracking is taking place. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is unauthorized data collection" is displayed in real time.

[1425] Specific examples

[1426] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's device will display a warning message stating, "This site is engaging in unauthorized data collection."

[1427] Prompt Sentence Examples

[1428] "Check if this site is engaging in unauthorized data collection or tracking."

[1429] Through this system, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling users to enjoy a safer and more reliable information environment.

[1430] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1431] News article credibility assessment

[1432] Step 1:

[1433] When a user views a news article, they click on the news article's URL.

[1434] Specific actions

[1435] When a user clicks on a news article link in an internet browser, JavaScript runs and retrieves the URL of the news article.

[1436] input

[1437] The URL of the news article the user clicked on.

[1438] output

[1439] The URL of the news article is retrieved on the user's device.

[1440] Step 2:

[1441] The device sends the URL of the news article to the server.

[1442] Specific actions

[1443] The JavaScript code sends the obtained URL to the backend server as a POST request.

[1444] input

[1445] The URL of the news article.

[1446] output

[1447] The URL of the news article is sent to the server.

[1448] Step 3:

[1449] The server retrieves the text and metadata of the news article based on the received URL.

[1450] Specific actions

[1451] The server uses Python's requests library to send requests to news sites, and extracts the text and metadata from the HTML using BeautifulSoup or similar.

[1452] input

[1453] The URL of the news article.

[1454] output

[1455] News article text and metadata.

[1456] Step 4:

[1457] The server sends the acquired news article text and metadata to the generative AI model, which generates an analysis prompt.

[1458] Specific actions

[1459] The server inputs the text and metadata of the news article into a generative AI (e.g., GPT-4) and generates a prompt saying, "Please rate the credibility of this article."

[1460] input

[1461] News article text and metadata.

[1462] output

[1463] The generated analysis prompt.

[1464] Step 5:

[1465] A generative AI model analyzes news articles and calculates a credibility score.

[1466] Specific actions

[1467] The generative AI model analyzes the text and metadata of the input news article and calculates a credibility score.

[1468] input

[1469] Parse prompts, news article body text and metadata.

[1470] output

[1471] Reliability scores and analysis results.

[1472] Step 6:

[1473] The server transmits the calculated reliability score and the analysis result to the user terminal.

[1474] Specific actions

[1475] The server returns the reliability score and analysis results in JSON format to the user device via the REST API set at the endpoint.

[1476] input

[1477] Reliability scores and analysis results.

[1478] output

[1479] The reliability score and analysis results sent to the user's device.

[1480] Step 7:

[1481] Based on the information received by the terminal, a warning message is displayed on the user interface.

[1482] Specific actions

[1483] The user interface (HTML, CSS, JavaScript) receives the reliability score and analysis results from the server and displays a warning message such as "This news article is not reliable" as a popup or banner.

[1484] input

[1485] Reliability scores and analysis results.

[1486] output

[1487] The warning message displayed in the user interface.

[1488] Detecting edited video and image data

[1489] Step 1:

[1490] A user uploads a video or image file onto the Internet.

[1491] Specific actions

[1492] The user interface displays a file selection dialog and the user selects a file.

[1493] input

[1494] Video or image files selected by the user.

[1495] output

[1496] The selected file will be uploaded to the system.

[1497] Step 2:

[1498] The terminal transmits the selected video or image data to the server.

[1499] Specific actions

[1500] Use the JavaScript FormData object to send the selected file to the server via a POST request.

[1501] input

[1502] Video or image files.

[1503] output

[1504] Video or image files sent to the server.

[1505] Step 3:

[1506] The server sends the received video or image data to the generative AI model, which generates an analysis prompt.

[1507] Specific actions

[1508] The server inputs the selected file data into a generation AI (e.g., DALL-E) and generates a prompt saying, "Please detect any editing traces in this image."

[1509] input

[1510] Uploaded video or image files.

[1511] output

[1512] The generated analysis prompt.

[1513] Step 4:

[1514] The generative AI model checks the video or image data for signs of editing and sends the results back to the server.

[1515] Specific actions

[1516] The AI ​​detects unnatural changes, color inconsistencies, and shadow discontinuities in the image and sends the analysis results back to the server.

[1517] input

[1518] Video or image file, analysis prompt.

[1519] output

[1520] Analysis results including traces of editing.

[1521] Step 5:

[1522] The server transmits the analysis results to the user terminal.

[1523] Specific actions

[1524] The server returns the analysis results in JSON format to the user device via the REST API set at the endpoint.

[1525] input

[1526] Analysis results including traces of editing.

[1527] output

[1528] Analysis results sent to the user's device.

[1529] Step 6:

[1530] Based on the analysis results, the device displays a warning message on the user interface.

[1531] Specific actions

[1532] The user interface receives the analysis results and displays a warning message such as "This footage may be faked" as a pop-up or banner.

[1533] input

[1534] Analysis results including traces of editing.

[1535] output

[1536] The warning message displayed in the user interface.

[1537] Social Media Monitoring

[1538] Step 1:

[1539] The browser extension monitors your online activity as you browse social media sites.

[1540] Specific actions

[1541] The browser extension collects the URLs of the sites you visit and the scripts those sites run.

[1542] input

[1543] Information about the social media sites you visit.

[1544] output

[1545] Site Information Collected.

[1546] Step 2:

[1547] The terminal sends the collected site information to the server.

[1548] Specific actions

[1549] The browser extension sends the collected data in JSON format to the server via a POST request.

[1550] input

[1551] Site Information Collected.

[1552] output

[1553] Site information sent to the server.

[1554] Step 3:

[1555] The server analyzes the site information it receives to determine whether unauthorized data collection or tracking is occurring.

[1556] Specific actions

[1557] The server analyzes the collected data against industry-standard security rules (for example, the OWASP threat list).

[1558] input

[1559] Received site information.

[1560] output

[1561] Verification of fraudulent data collection or tracking.

[1562] Step 4:

[1563] The server transmits the determination result to the user terminal.

[1564] Specific actions

[1565] The server returns the result of the judgment in JSON format to the user device via the REST API set at the endpoint.

[1566] input

[1567] Verification of fraudulent data collection or tracking.

[1568] output

[1569] The judgment result sent to the user terminal.

[1570] Step 5:

[1571] Based on the judgment result, the terminal displays a warning message on the user interface.

[1572] Specific actions

[1573] The user interface receives the judgment result and displays a warning message such as "This site is collecting data illegally" as a pop-up or banner.

[1574] input

[1575] Judgment result.

[1576] output

[1577] The warning message displayed in the user interface.

[1578] (Application example 1)

[1579] Next, a description will be given of Application Example 1. 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."

[1580] The decline in the reliability of information on the Internet and the resulting spread of misinformation and fake news have become a social problem. Furthermore, the increasing number of edited videos and images, as well as the increasing number of illegal data collection and tracking practices on social media, are making it difficult for users to determine the authenticity of information. These problems must be resolved quickly and accurately, and users must be provided with reliable information.

[1581] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1582] In this invention, the server includes means for collecting data via a network and acquiring news article text and metadata, means for analyzing the acquired text and metadata and calculating a reliability score for the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for receiving and analyzing video or image data and detecting editing and unauthorized manipulation, and means for monitoring and determining unauthorized data collection and tracking on social media sites and displaying a warning to the user. This allows users to quickly evaluate the reliability of information on the Internet and be protected from unauthorized data collection.

[1583] A "network" is a collection of computers and communication devices that allow information and data to be sent and received.

[1584] "Data" refers to the information being analyzed, such as news articles, videos, images, and social media information.

[1585] A "news article" is written information such as a news report posted on the Internet.

[1586] "Body" refers to the main content of a news article.

[1587] "Metadata" is supplemental information and attribute data associated with a news article.

[1588] A "trust score" is a numerical assessment of the trustworthiness of a particular article or piece of information.

[1589] A "warning" is a message that is displayed to the user to call attention to the situation.

[1590] "Video" is data that contains moving visual information.

[1591] An "image" is data containing still visual information.

[1592] "Editing evidence" refers to evidence of changes that indicate an image or video has been modified.

[1593] "Social media" is an online platform for sharing information between users.

[1594] "Unauthorized data collection" refers to the illegal act of collecting and using user information without permission.

[1595] "Tracking" is the act of following a user's online activities.

[1596] A "user terminal" is a device used to view information and receive alerts.

[1597] A "generative AI model" is a machine learning model for performing natural language processing and data analysis.

[1598] A "prompt sentence" is a sentence used to input instructions or questions into a generative AI model.

[1599] This invention provides a system for evaluating the reliability of information on the Internet. The system analyzes news articles, video and image data, and information on social media, and conveys the reliability to users.

[1600] Program Generation

[1601] The following describes an outline of a program and its processing for realizing the present invention.

[1602] Hardware / Software used

[1603] Hardware: User's smartphone, server

[1604] Software: Mobile apps, browser extensions, generative AI models (e.g., GPT-4), data analysis tools

[1605] News article credibility assessment

[1606] When a user views a news article on their smartphone, the URL of that news article is automatically sent to a server. A generative AI model (e.g., GPT-4) on the server retrieves and analyzes the news article text and metadata from the URL. Based on the analysis results, a reliability score is calculated and sent to the user's smartphone. If the reliability score is low, the user's smartphone displays a warning message such as "This news article is unreliable."

[1607] Examples:

[1608] When a user clicks on a news article shared on a social networking site, the URL is sent to a server, which analyzes the news article and, if the reliability score is low, displays a warning message on the user's smartphone saying, "This news article is unreliable."

[1609] Prompt for the generative AI model:

[1610] News article URL: [example.com / news123]

[1611] Rate the article's credibility and calculate a credibility score, and explain why it contains exaggerated or unreliable content.

[1612] Detecting traces of editing of video and image data

[1613] When a user views a video or image file on their smartphone, the file is automatically uploaded to a server. A generative AI model on the server analyzes the video or image and detects any signs of editing. If there is a problem with its authenticity, that information is sent to the user's smartphone, warning them that "this video may be faked."

[1614] Examples:

[1615] When a user downloads a video shared on a social networking site and uploads it to the system, the server analyzes the video, and if any fraud is detected, a warning message will appear on the smartphone saying, "This video may be faked."

[1616] Social Media Monitoring

[1617] When a user visits a social media site in their browser, the browser extension monitors the user's online activity and sends the site information to a server. An analysis tool on the server identifies any unauthorized data collection or tracking attempts and sends the results to the user's smartphone. If any unauthorized activity is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed.

[1618] Examples:

[1619] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's smartphone will display a warning message stating, "This site is engaging in unauthorized data collection."

[1620] In this way, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling them to enjoy a safer and more reliable information environment.

[1621] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1622] Program flow:

[1623] News article credibility assessment

[1624] Step 1:

[1625] The user terminal inputs the URL of the news article.

[1626] The entered URL is sent from the user's smartphone to the server.

[1627] Step 2:

[1628] The server retrieves the text and metadata of the news article based on the received URL.

[1629] The server uses a web scraping tool to retrieve data from the URL.

[1630] Step 3:

[1631] The server parses the retrieved text and metadata.

[1632] A generative AI model (e.g., GPT-4) is used to calculate a confidence score for the text.

[1633] Step 4:

[1634] The server transmits the calculated reliability score to the user terminal.

[1635] The reliability score and analysis results are displayed on the user's smartphone.

[1636] Step 5:

[1637] If the user terminal has a low reliability score, a warning message is displayed.

[1638] Users may see a message that reads, "This news article is not reliable."

[1639] Detecting traces of editing of video and image data

[1640] Step 1:

[1641] The user uploads a video or image to the device.

[1642] The uploaded data is automatically sent to the server.

[1643] Step 2:

[1644] The server analyzes the received video or image data.

[1645] Generative AI models are used to detect edits (e.g., unnatural color tones and shadows).

[1646] Step 3:

[1647] The server transmits the analysis results to the user terminal.

[1648] A warning message appears on the user's smartphone saying, "This video may be faked."

[1649] Step 4:

[1650] The user terminal displays a warning message and presents the analysis results.

[1651] Users can see the specific locations and reasons for unauthorized edits.

[1652] Social Media Monitoring

[1653] Step 1:

[1654] A user browses a social media site.

[1655] Browser extensions monitor your online activity and collect site information.

[1656] Step 2:

[1657] The server analyzes the collected site information.

[1658] Generative AI models identify potential fraudulent data collection and tracking.

[1659] Step 3:

[1660] The server transmits the determination result to the user terminal.

[1661] A warning message appears on the user's smartphone saying, "This site is collecting data illegally."

[1662] Step 4:

[1663] The user terminal displays a warning message and provides details of the fraudulent activity.

[1664] Users will be given a detailed explanation of what fraudulent activity is taking place.

[1665] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1666] ---

[1667] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[1668] Evaluating the credibility of news articles and using sentiment engines

[1669] When a user views a news article online, the user's device sends the news article's URL to a server. The server retrieves the news article's text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the text content, source reliability, past history, etc. The user's emotion engine then recognizes the user's emotional state (excited, angry, calm, etc.). The calculated reliability score and analysis results are sent to the user's device, and the way the warning is displayed is adjusted based on the user's emotional state detected by the emotion engine. When a warning message such as "This news article is unreliable" is displayed, if the user is angry, the warning will be displayed more gently, and if the user is calm, a more direct warning will be displayed.

[1670] Specific examples

[1671] When a user clicks on a news article shared on a social networking site, the URL is sent to the server. The server analyzes the news article, and if the reliability score is low, a warning message is displayed on the user's device. At the same time, the emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a gentle warning if the user is excited.

[1672] Detecting edited video and image data and using emotion engines

[1673] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generation AI checks the video or image for signs of editing and identifies any unauthorized changes. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. The warning, "This video may be faked," is adjusted according to the user's state.

[1674] Specific examples

[1675] Users download videos posted on social media and upload them to the system. The server analyzes the video to determine whether there are any signs of unauthorized editing. If unauthorized editing is detected, the emotion engine recognizes the user's emotions and displays a gentle warning message if the user is feeling angry.

[1676] Social media monitoring and sentiment engine utilization

[1677] When a user visits a social media site in a browser, a browser extension is used to monitor the user's online activity. The user's device sends information about the site being accessed to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The result of the determination is sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message is displayed encouraging them to stay calm. The warning, "This site is engaging in unauthorized data collection," is adjusted according to the user's state.

[1678] Specific examples

[1679] As a user browses social media sites, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, an emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a more gentle message if the user is impatient.

[1680] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into account, providing more effective and responsive warnings. This allows users to enjoy a safer and more reliable information environment.

[1681] ---

[1682] This concludes the description of the embodiment of the present invention. This system can more effectively ensure the accuracy of information and provide appropriate warnings to users by evaluating the reliability of news articles, detecting edits in video and image data, and monitoring social media, as well as taking into account the emotional state of the user.

[1683] The processing flow will be explained below.

[1684] Evaluating the credibility of news articles and using sentiment engines

[1685] Step 1:

[1686] A user clicks on a news article link, and the user's device sends the news article URL to the server.

[1687] Step 2:

[1688] The server retrieves the HTML data of the news article from the received URL.

[1689] Step 3:

[1690] The server extracts the news article text and metadata from the HTML data it has obtained.

[1691] Step 4:

[1692] The generation AI placed on the server analyzes the text of news articles using natural language processing (NLP) technology and evaluates the content of the article, the reliability of the source, past history, etc.

[1693] Step 5:

[1694] The server calculates a reliability score based on the analysis results.

[1695] Step 6:

[1696] The server queries the user's emotion engine for the user's most recent emotional state.

[1697] Step 7:

[1698] The user's device activates an emotion engine to detect the user's emotional state (excitement, anger, calmness, etc.).

[1699] Step 8:

[1700] The emotion engine transmits the user's emotional state to the server.

[1701] Step 9:

[1702] The server adjusts how the warning message is displayed depending on the confidence score and the detected emotional state.

[1703] Step 10:

[1704] The server sends a reliability score and a warning message to the user's device.

[1705] Step 11:

[1706] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[1707] Detecting edited video and image data and using emotion engines

[1708] Step 1:

[1709] A user uploads video or image data to the system.

[1710] Step 2:

[1711] The user's terminal transmits the uploaded file data to the server.

[1712] Step 3:

[1713] The server receives the file data.

[1714] Step 4:

[1715] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[1716] Step 5:

[1717] The server creates a warning message based on the analysis results.

[1718] Step 6:

[1719] The server queries the user's emotion engine for the emotional state.

[1720] Step 7:

[1721] The user's terminal activates an emotion engine to detect the user's emotional state.

[1722] Step 8:

[1723] The emotion engine transmits the user's emotional state to the server.

[1724] Step 9:

[1725] The server tailors the warning message based on the analysis results and emotional state.

[1726] Step 10:

[1727] The server sends the analysis results and warning messages to the user's terminal.

[1728] Step 11:

[1729] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[1730] Social media monitoring and sentiment engine utilization

[1731] Step 1:

[1732] A user browses to a social media site.

[1733] Step 2:

[1734] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[1735] Step 3:

[1736] The user's terminal transmits the collected site information to the server.

[1737] Step 4:

[1738] The server analyzes the received site information.

[1739] Step 5:

[1740] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[1741] Step 6:

[1742] The server generates a verdict and a warning message.

[1743] Step 7:

[1744] The server queries the user's emotion engine for the emotional state.

[1745] Step 8:

[1746] The user's terminal activates an emotion engine to detect the user's emotional state.

[1747] Step 9:

[1748] The emotion engine transmits the user's emotional state to the server.

[1749] Step 10:

[1750] The server adjusts the warning message based on the result of the assessment and the emotional state.

[1751] Step 11:

[1752] The server sends the judgment result and a warning message to the user's terminal.

[1753] Step 12:

[1754] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[1755] ---

[1756] In this way, by taking into account the user's emotional state in addition to assessing the reliability of news articles, detecting traces of editing in video and image data, and monitoring social media, we can more effectively ensure the accuracy of information and provide appropriate warnings to users.

[1757] Example 2

[1758] Next, a description will be given of Example 2. 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."

[1759] The internet is flooded with information, often containing unreliable news articles and fraudulently manipulated video and images. Social media sites also sometimes conduct fraudulent data collection and tracking. This information not only causes confusion and anxiety among users, but can also lead to social unrest. Therefore, it is important to evaluate the reliability of this information and provide appropriate warnings to users. However, conventional systems often display uniform warnings without considering the user's emotional state, which can detract from the user experience.

[1760] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1761] In this invention, the server includes means for acquiring news article text and metadata, means for analyzing the acquired text and metadata and using a generative AI model to calculate a reliability score for the information, means for using an emotion engine to recognize the user's emotional state, means for adjusting a warning based on the calculated reliability score and the user's emotional state and sending the adjusted warning message to the user terminal, and means for displaying the adjusted warning message on the user terminal. This makes it possible to evaluate the reliability of news articles, video / image data, and social media site information, and to provide flexible warnings that take the user's emotional state into consideration.

[1762] "Means for obtaining news article text and metadata" refers to a system for collecting the text of news articles published on the Internet and related metadata such as publication date, author information, and URL.

[1763] "Means for using generative AI models" refers to a mechanism for using generative artificial intelligence models to analyze collected data and output the results.

[1764] An "emotion engine for recognizing the user's emotional state" is an engine that can identify and analyze the user's emotions, and is a mechanism for determining whether the user is excited, angry, calm, etc.

[1765] The "means for calculating the reliability score of information" is a system that analyzes the content and metadata of news articles, video and image data, and expresses the reliability of that information as a quantified score.

[1766] The "means for adjusting and sending a warning to a user terminal" is a mechanism for adjusting the content and display method of a warning message based on the calculated reliability score and the recognized emotional state of the user, and sending it to the user's terminal.

[1767] The "means for displaying the adjusted warning message on the user terminal" is a mechanism for displaying the adjusted warning message sent from the server on the user terminal.

[1768] "Means for receiving video or image data" refers to a mechanism for receiving video or image data uploaded by a user.

[1769] "Means for identifying traces of unauthorized editing" refers to a mechanism for analyzing received video or image data and checking for traces of editing or manipulation.

[1770] "Means for collecting site information" refers to a mechanism for collecting information on the social media sites that users access.

[1771] "Means for determining attempts at unauthorized data collection or tracking" refers to a mechanism for analyzing collected site information to determine whether the site is attempting unauthorized data collection or tracking of users.

[1772] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[1773] Evaluating the credibility of news articles and using sentiment engines

[1774] When a user browses a news article online, the user's device sends the URL of the news article to a server. The server retrieves the news article's text and metadata from the URL. A generative AI deployed on the server analyzes the news article and calculates a credibility score based on the text's content, the credibility of the source, and past history. Specifically, a generative AI model (such as GPT-4 for natural language processing) is used to analyze the text and assign a credibility score. An emotion engine (such as IBM Watson's Tone Analyzer) then recognizes the user's emotional state. The calculated credibility score and analysis results are sent to the user's device, and the display of the warning is adjusted based on the user's emotional state detected by the emotion engine. For example, when a warning message stating "This news article is unreliable" is displayed to a user, if the user is angry, the warning will be displayed more gently, whereas if the user is calm, a more direct warning will be displayed.

[1775] Example prompt sentence:

[1776] "Please parse the following news article URL and calculate a credibility score: [news article URL]"

[1777] Detecting edited video and image data and using emotion engines

[1778] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generated AI checks the video or image for signs of editing and identifies any unauthorized changes. For example, it uses Adobe Photoshop's image analysis function to detect editing. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. For example, the warning "This video may be faked" is adjusted according to the user's state.

[1779] Example prompt sentence:

[1780] "Analyze this video file to check for any signs of unauthorized editing: [video file path]"

[1781] Social media monitoring and sentiment engine utilization

[1782] When a user browses social media sites in a browser, their online activity is monitored using a browser extension. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. For example, this uses the monitoring function of Ghostery. The results of this determination are sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message urging them to stay calm is displayed. For example, the warning "This site is engaging in unauthorized data collection" is adjusted according to the user's state.

[1783] Example prompt sentence:

[1784] "Check if this social media site is illegally collecting data: [site URL]"

[1785] The goal of this system is to evaluate the reliability of news articles, video and image data, and social media information, helping users easily determine the accuracy of the information, while also taking into account the user's emotional state to provide more effective warnings.

[1786] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1787] Evaluating the credibility of news articles and using sentiment engines

[1788] Step 1:

[1789] A user clicks on a news article URL.

[1790] (Input) The URL the user clicks on.

[1791] (Output) URL information clicked on the user's device.

[1792] Users click on the URL of the news article they want to view in their web browser.

[1793] Step 2:

[1794] The device sends the URL information to the server.

[1795] (Input) The URL of the news article that was clicked.

[1796] (Output) The URL information sent to the server.

[1797] The user's terminal sends the URL of the clicked news article to the server as an HTTP request.

[1798] Step 3:

[1799] The server retrieves the news article text and metadata.

[1800] (Input) The URL of the received news article.

[1801] (Output) The news article text and metadata.

[1802] The server uses the received URL to obtain the text of the news article and metadata (publication date, author name, etc.) from the news site using scraping or an API.

[1803] Step 4:

[1804] The server instructs the generated AI to analyze.

[1805] (Input) News article text and metadata.

[1806] (Output) Analysis instructions to the generating AI.

[1807] The server sends the text and metadata of the retrieved news article as input to the generation AI, instructing it to analyze it.

[1808] Step 5:

[1809] The generative AI calculates a reliability score.

[1810] (Input) News article text and metadata.

[1811] (Output) The confidence score.

[1812] The generative AI (e.g., GPT-4) analyzes the text and metadata of the received news article and calculates a credibility score based on the content of the text, the reliability of the source, past history, etc.

[1813] Step 6:

[1814] An emotion engine recognizes the user's emotional state.

[1815] (Input) User behavior data (viewing time, scrolling speed, etc.).

[1816] (Output) The user's emotional state.

[1817] An emotion engine (e.g., IBM Watson Tone Analyzer) analyzes the user's behavioral data and recognizes the user's current emotional state (excited, angry, calm, etc.).

[1818] Step 7:

[1819] The server sends the analysis results and warning messages to the terminal.

[1820] (Input) Confidence score and user's emotional state.

[1821] (Output) The warning message.

[1822] The server adjusts the content of the warning message based on the reliability score calculated by the generation AI and the user's emotional state recognized by the emotion engine, and sends it to the user's device.

[1823] Step 8:

[1824] The terminal displays a warning message to the user.

[1825] (Input) The warning message received from the server.

[1826] (Output) The warning message to be displayed.

[1827] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the news article.

[1828] Detecting edited video and image data and using emotion engines

[1829] Step 1:

[1830] A user uploads video or image data to the system.

[1831] (Input) Video or image file.

[1832] (Output) The data uploaded to the system.

[1833] Users upload video or image files to the system via the Internet.

[1834] Step 2:

[1835] The device sends the data to the server.

[1836] (Input) Uploaded video or image data.

[1837] (Output) The data sent to the server.

[1838] The user's terminal transmits the uploaded video or image data to the server.

[1839] Step 3:

[1840] The server instructs the generated AI to analyze.

[1841] (Input) Received video or image data.

[1842] (Output) Analysis instructions to the generating AI.

[1843] The server sends the received data to the generation AI and instructs it to analyze it.

[1844] Step 4:

[1845] Generative AI checks for signs of editing in videos and images.

[1846] (Input) Video or image data.

[1847] (Output) Analysis results for edit traces.

[1848] The generating AI analyzes the received video and image data and checks for signs of editing or unauthorized processing.

[1849] Step 5:

[1850] An emotion engine recognizes the user's emotional state.

[1851] (Input) User behavior data (viewing time, reaction speed, etc.).

[1852] (Output) The user's emotional state.

[1853] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[1854] Step 6:

[1855] The server sends the analysis results and warning messages to the terminal.

[1856] (Input) Analysis results of editing traces and the user's emotional state.

[1857] (Output) The warning message.

[1858] The server adjusts the content of the warning message based on the analysis results and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[1859] Step 7:

[1860] The terminal displays a warning message to the user.

[1861] (Input) The warning message received from the server.

[1862] (Output) The warning message to be displayed.

[1863] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the video or image.

[1864] Social media monitoring and sentiment engine utilization

[1865] Step 1:

[1866] A user browses a social media site.

[1867] (Input) Social media site URL.

[1868] (Output) Site browsing information on the user's device.

[1869] Users use a web browser to browse social media sites.

[1870] Step 2:

[1871] The site information that the terminal is accessing is sent to the server.

[1872] (Input) Information about the site you are viewing.

[1873] (Output) Site information sent to the server.

[1874] The browser extension collects information about the social media sites you visit and sends it to a server.

[1875] Step 3:

[1876] The server analyzes the site information.

[1877] (Input) Received site information.

[1878] (Output) Site analysis results.

[1879] The server analyzes the received information from the social media site and evaluates its content.

[1880] Step 4:

[1881] The server determines whether or not unauthorized data collection has occurred.

[1882] (Input) Site analysis results.

[1883] (Output) The result of the invalid data collection determination.

[1884] The server uses the site information to determine whether unauthorized data collection or tracking is occurring.

[1885] Step 5:

[1886] An emotion engine recognizes the user's emotional state.

[1887] (Input) User behavior data (viewing time, reaction speed, etc.).

[1888] (Output) The user's emotional state.

[1889] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[1890] Step 6:

[1891] The server sends the judgment result and a warning message to the terminal.

[1892] (Input) The result of the fraudulent data collection and the user's emotional state.

[1893] (Output) The warning message.

[1894] The server adjusts the content of the warning message based on the judgment result and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[1895] Step 7:

[1896] The terminal displays a warning message to the user.

[1897] (Input) The warning message received from the server.

[1898] (Output) The warning message to be displayed.

[1899] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning about the trustworthiness of the social media site.

[1900] (Application example 2)

[1901] Next, a description will be given of Application Example 2. 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."

[1902] In today's Internet environment, there is a lot of false information and incorrect data, making it difficult for users to determine whether or not to trust this information. Displaying appropriate warning messages based on the user's emotional state is also problematic. Conventional systems display uniform warning messages that ignore the user's emotional state, which can prevent users from receiving warnings appropriately.

[1903] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring the text and metadata of news articles, means for analyzing the acquired text and metadata and calculating a reliability score of the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for recognizing the user's emotional state and adjusting the display of the warning message based on the emotional state, and means for analyzing the URL and metadata of video content, calculating a reliability score, and displaying a warning message. This allows users to quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into consideration, providing more effective and responsive warnings.

[1904] A "network" is an infrastructure that allows multiple computers and devices to communicate with each other.

[1905] "Data" is a coded set of letters and numbers that represent specific information.

[1906] A "news article" is a piece of writing that reports social, economic, political, or cultural events or information.

[1907] "Body" is the text portion that constitutes the main content of a news article.

[1908] "Metadata" is data that describes information and attributes about data.

[1909] A "trustworthiness score" is an evaluation index that quantifies the accuracy and reliability of information.

[1910] A "user terminal" is a device used by a user to obtain and operate information.

[1911] A "warning" is a message intended to inform the user of a potential risk or problem.

[1912] "Emotional state" refers to the user's current psychological response or emotional state.

[1913] A "warning message" is textual or graphical information displayed to inform the user of a warning.

[1914] "Video content" refers to media data that is viewed as moving images.

[1915] A "URL" is an address that identifies the location of a web page or other Internet resource.

[1916] "Analysis" is the process of examining data in detail to reveal its structure and meaning.

[1917] "Unauthorized editing" is any deliberate alteration made to a footage or image that is false or misleading.

[1918] "Online activities" refer to the various operations and actions that users perform using the Internet.

[1919] "Data collection" is the act of collecting and storing information for a specific purpose.

[1920] "Tracking" refers to any technology or process that follows or records a user's online actions or movements.

[1921] A "generative AI model" is a model that is trained to solve a specific problem using artificial intelligence techniques.

[1922] A "prompt" is a text instruction entered into a generative AI model.

[1923] This invention relates to a system that evaluates the reliability of information in content distribution services and displays appropriate warnings according to the user's emotional state. This system provides users with a safe and reliable information environment by monitoring news articles, video content, video and image data, and social media sites.

[1924] Key Components

[1925] The main components of this system are as follows:

[1926] 1. Data collection method: A module for collecting URLs and metadata of news articles and video content. This includes a mechanism for sending data from the user's device to the server.

[1927] 2. Analysis: Analyzes the text and metadata of incoming news articles and video content and calculates a credibility score for the information, using a generative AI model.

[1928] 3. Emotion Recognition Means: A module that recognizes the user's current emotional state. This includes emotion recognition software.

[1929] 4. Warning display means: A module for displaying appropriate warning messages to users based on their confidence scores and emotional states.

[1930] Hardware and software used

[1931] Hardware: User devices (smartphones, tablets, PCs, smart glasses), servers.

[1932] software:

[1933] Emotion recognition software (e.g., Microsoft Azure Emotion API)

[1934] Video analysis software (e.g., Google Cloud Video Intelligence API)

[1935] Front-end application for displaying alerts (e.g. React Native)

[1936] Data processing and data calculation

[1937] The server retrieves the URLs and metadata of news articles and video content and analyzes them using a generative AI model. Specifically, it uses the Google Cloud Video Intelligence API to analyze the video content and calculate a reliability score. For emotion recognition, it uses the Microsoft Azure Emotion API to detect the user's emotional state. The calculated reliability score and analysis results are sent to the user's device, and the display of warning messages is adjusted based on the user's emotional state.

[1938] Specific examples

[1939] When a user watches a news video on a content delivery service, the video's URL and metadata are sent to a server. The server analyzes the video using the Google Cloud Video Intelligence API, and if the video is deemed to have a low reliability score, it uses emotion recognition software to assess the user's emotional state. For example, if the user is excited, the system displays a mild-toned warning message saying, "Please watch with caution. This video may contain unreliable information." If the user is calm, a more direct warning message is displayed.

[1940] Prompt Sentence Examples

[1941] "This video may be unreliable. Please watch with caution. This video may contain unreliable information."

[1942] This allows users to quickly evaluate the accuracy of information and use the internet safely. Furthermore, the emotion engine optimizes warning messages, allowing users to receive information appropriately.

[1943] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1944] Step 1:

[1945] When a user watches video content on a content distribution service, the URL and metadata of that video are sent from the user's device to the server. Specifically, when the user presses the watch button, the URL and metadata are collected and sent to the server via the network. This allows the server to obtain the video information to analyze.

[1946] Step 2:

[1947] The server passes the received video URL and metadata to the Google Cloud Video Intelligence API, which analyzes the video content. The analysis calculates a credibility score based on the video's content, source credibility, past history, etc. The input for this step is the video URL and metadata, and the output is a credibility score.

[1948] Step 3:

[1949] The server sends the calculated reliability score to the user's device. At the same time, emotion recognition software (Microsoft Azure Emotion API) is launched on the user's device and recognizes the user's current emotional state through the camera. The inputs of this step are the reliability score and the user's camera image, and the output is the user's emotional state data.

[1950] Step 4:

[1951] The emotion recognition software analyzes the user's emotional state and sends the results to the server. Specifically, it identifies emotions from the user's facial expressions and vocal tone and returns the analysis results to the server. The input for this step is the camera footage, and the output is the recognized emotional state data.

[1952] Step 5:

[1953] The server adjusts the content and tone of the warning message based on the reliability score and the user's emotional state. For example, if the reliability score is low and the user is in an excited state, the warning message will read, "Please watch with caution. This video may contain unreliable information." The inputs of this step are the reliability score and the user's emotional state, and the output is the adjusted warning message.

[1954] Step 6:

[1955] Finally, the adjusted warning message is sent to the user terminal and displayed on the video content. The user can check the warning message while watching the video. The input of this step is the adjusted warning message, and the output is the display of the warning message to the user.

[1956] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1957] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1958] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1959] [Fourth embodiment]

[1960] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1961] 7, a 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.

[1962] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1963] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1964] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1965] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1966] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1967] The control object 443 includes a display device, LEDs in the eyes, and motors for driving 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1968] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1969] The specific processing program 56 is an example of a "program" according to the technology of the present 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.

[1970] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1971] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. 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 process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1972] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1973] ---

[1974] The present invention relates to a system for ensuring the accuracy of information on the Internet, and can be implemented as follows: The system of the present invention aims to evaluate and analyze news articles, video and image data, and information through social media to ensure their reliability and accuracy.

[1975] News article credibility assessment

[1976] When a user views a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the news article text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the content of the text, the reliability of the source, past history, etc. The calculated reliability score and analysis results are then sent to the user's device, and a warning message such as "This news article has low reliability" is displayed on the user's device.

[1977] Specific examples

[1978] When a user clicks on a news article shared on a social networking site, the URL is sent to the server, which analyzes the news article and displays a warning message on the user's device if the article has a low reliability score. In this case, the user can confirm the reliability of the news article.

[1979] Detecting edited video and image data

[1980] When a user browses video or image files online, they can upload them to the system. The user's device sends the video or image data to the server, which analyzes it. The server's AI then checks the video or image for signs of editing and identifies unnatural changes, such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message is displayed, such as "This video may be faked."

[1981] Specific examples

[1982] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If any unauthorized editing is detected, a warning is displayed on the user's device.

[1983] Social Media Monitoring

[1984] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed in real time.

[1985] Specific examples

[1986] As users browse social media sites, the browser extension monitors the site's scripts and requests, and if the server determines that the site is engaging in unauthorized data collection, a warning message is displayed on the user's device.

[1987] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection, allowing them to enjoy a safer and more reliable information environment.

[1988] ---

[1989] This concludes the description of the present invention, which provides concrete steps to improve the reliability of information in news articles, video and image data, and social media monitoring.

[1990] The processing flow will be explained below.

[1991] News article credibility assessment

[1992] Step 1:

[1993] A user clicks on a link to a news article. The user's device sends this link information to the server.

[1994] Step 2:

[1995] The server retrieves the HTML data of the news article from the received URL.

[1996] Step 3:

[1997] The server extracts the news article text and metadata from the HTML data it has obtained.

[1998] Step 4:

[1999] The generation AI placed on the server analyzes the text of the news article using natural language processing (NLP) technology and evaluates the article's content, source reliability, past history, etc.

[2000] Step 5:

[2001] The server calculates a reliability score based on the analysis results.

[2002] Step 6:

[2003] The server sends the calculated reliability score and a summary of the analysis results to the user's device.

[2004] Step 7:

[2005] The analysis results received by the user's device are displayed, and if necessary, a warning message such as "This news article is unreliable" is displayed.

[2006] Detecting edited video and image data

[2007] Step 1:

[2008] A user uploads video or image data to the system.

[2009] Step 2:

[2010] The user's terminal transmits the uploaded file data to the server.

[2011] Step 3:

[2012] The server receives the file data.

[2013] Step 4:

[2014] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[2015] Step 5:

[2016] The server sends the analysis results to the user's device.

[2017] Step 6:

[2018] The analysis results received by the user's device are displayed, along with a warning message such as "This video may be faked."

[2019] Social Media Monitoring

[2020] Step 1:

[2021] A user browses to a social media site.

[2022] Step 2:

[2023] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[2024] Step 3:

[2025] The user's terminal transmits the collected site information to the server.

[2026] Step 4:

[2027] The server analyzes the received site information.

[2028] Step 5:

[2029] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[2030] Step 6:

[2031] The server sends the analysis results to the user's device.

[2032] Step 7:

[2033] The analysis results received by the user's device are displayed, and a warning message such as "This site is collecting data illegally" is displayed in real time.

[2034] ---

[2035] These are the specific processing steps for assessing the credibility of news articles, detecting edited video and image data, and monitoring social media. By following these steps, the system can efficiently evaluate information and provide necessary warnings to users.

[2036] Example 1

[2037] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2038] The problem is that it is difficult for users to easily determine the authenticity of news articles, video and image data, and information on social media, which is flooding the internet. This can lead to the spread of false information and the violation of privacy through the unauthorized collection of data.

[2039] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[2040] In this invention, the server includes: means for transmitting a specific URL of a news article from a user's device to the server; means for receiving the transmitted URL at the server and acquiring the text and metadata of the news article; means for analyzing the acquired text and metadata using a generative AI model located on the server to calculate a reliability score of the information; means for transmitting the calculated reliability score and the analysis result to the user's device and displaying them as a warning message; means for receiving video or image data uploaded from the user's device and transmitting the received video or image data to the server to detect editing and unauthorized manipulation; means for analyzing the transmitted data using a generative AI model located on the server to identify traces of unauthorized editing; means for transmitting the analysis result to the user's device and displaying a warning message; means for monitoring the user's online activity using a browser extension and collecting information about the sites accessed; means for transmitting the collected site information to the server, analyzing it using the generative AI model located on the server, and determining attempts of unauthorized data collection or tracking; and means for transmitting the determination result to the user's device and displaying it as a warning message in real time. This allows users to quickly and easily evaluate the reliability of news articles, video and image data, and information on social media and protect themselves from unauthorized data collection.

[2041] A "news article" is an article or news report published on the Internet, including the text and metadata such as the title, author, and publication date.

[2042] "URL" is an abbreviation for Uniform Resource Locator, and is a character string that indicates the address of a specific resource on the Internet.

[2043] A "generative AI model" is an artificial intelligence model used to perform natural language processing and data analysis, such as a machine learning model that generates sentences and performs data analysis.

[2044] The "trustworthiness score" is a numerical representation of the reliability and accuracy of news articles, video and image data, social media information, etc., and is an indicator of their reliability.

[2045] "Analysis" is the process of using generative AI models to examine the content of data in detail and evaluate its content, the reliability of its sources, and evidence of editing.

[2046] A "warning message" is a message that displays important information or a warning to the user, and is particularly a warning about information that is unreliable or inaccurate.

[2047] "Video data" refers to digital data stored in a moving image format, including video clips and movies.

[2048] "Image data" refers to digital data stored in still image format, including photographs and drawings.

[2049] "Editing traces" are unnatural changes or inconsistencies that serve as evidence that video or image data has been edited or manipulated.

[2050] A "browser extension" is additional software that extends the functionality of a web browser and has the ability to monitor a user's online activities and collect specific site information.

[2051] "Unauthorized data collection" refers to the act of illegally collecting personal data or usage information without the user's consent.

[2052] "Tracking" refers to the act of tracking a user's website browsing behavior and collecting and analyzing that data.

[2053] This invention relates to a system that monitors news articles, video and image data, and social media to evaluate the accuracy and reliability of the information. This system uses a generative AI model to evaluate the reliability of news articles, detect editing traces in video and image data, and monitor social media.

[2054] News article credibility assessment

[2055] When a user browses a news article on the Internet, the user's device sends the URL of the news article to a server. The server retrieves the text and metadata of the news article based on the received URL. A generative AI model (e.g., GPT-4) deployed on the server analyzes the text and metadata of the retrieved news article and calculates a reliability score. The analysis result and reliability score are then sent back to the user's device and displayed as a warning message on the user interface.

[2056] Specific examples

[2057] When a user clicks on a news article shared on a social networking site, the URL is sent to a server. The server analyzes the news article and calculates a reliability score using a generative AI model. If the reliability score is low, a warning message stating "This news article is unreliable" is displayed on the user's device.

[2058] Prompt Sentence Examples

[2059] "Analyze the text of news articles and calculate a credibility score."

[2060] Detecting edited video and image data

[2061] When a user views video or image files on the internet, they upload these files to the system. The user's device sends the video or image data to a server, which analyzes the data. A generative AI model (e.g., DALL-E) located on the server checks the video or image for signs of editing and identifies unnatural changes such as color inconsistencies or shadow discontinuities. The analysis results are then sent back to the user's device, where a warning message such as "This video may be faked" is displayed.

[2062] Specific examples

[2063] Users download videos posted on social media and upload them to the system. The server analyzes the video and determines whether there are any signs of unauthorized editing. If unauthorized editing is detected, a warning message is displayed on the user's device stating, "This video may have been forged."

[2064] Prompt Sentence Examples

[2065] "Please detect any editing signs in this image and assess its authenticity."

[2066] Social Media Monitoring

[2067] When a user browses social media sites in a browser, the browser extension monitors the user's online activity. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether or not unauthorized data collection or tracking is taking place. The results of the assessment are sent to the user's device, and if a problem is detected, a warning message stating "This site is unauthorized data collection" is displayed in real time.

[2068] Specific examples

[2069] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's device will display a warning message stating, "This site is engaging in unauthorized data collection."

[2070] Prompt Sentence Examples

[2071] "Check if this site is engaging in unauthorized data collection or tracking."

[2072] Through this system, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling users to enjoy a safer and more reliable information environment.

[2073] The flow of the identification process in the first embodiment will be described with reference to FIG.

[2074] News article credibility assessment

[2075] Step 1:

[2076] When a user views a news article, they click on the news article's URL.

[2077] Specific actions

[2078] When a user clicks on a news article link in an internet browser, JavaScript runs and retrieves the URL of the news article.

[2079] input

[2080] The URL of the news article the user clicked on.

[2081] output

[2082] The URL of the news article is retrieved on the user's device.

[2083] Step 2:

[2084] The device sends the URL of the news article to the server.

[2085] Specific actions

[2086] The JavaScript code sends the obtained URL to the backend server as a POST request.

[2087] input

[2088] The URL of the news article.

[2089] output

[2090] The URL of the news article is sent to the server.

[2091] Step 3:

[2092] The server retrieves the text and metadata of the news article based on the received URL.

[2093] Specific actions

[2094] The server uses Python's requests library to send requests to news sites, and extracts the text and metadata from the HTML using BeautifulSoup or similar.

[2095] input

[2096] The URL of the news article.

[2097] output

[2098] News article text and metadata.

[2099] Step 4:

[2100] The server sends the acquired news article text and metadata to the generative AI model, which generates an analysis prompt.

[2101] Specific actions

[2102] The server inputs the text and metadata of the news article into a generative AI (e.g., GPT-4) and generates a prompt saying, "Please rate the credibility of this article."

[2103] input

[2104] News article text and metadata.

[2105] output

[2106] The generated analysis prompt.

[2107] Step 5:

[2108] A generative AI model analyzes news articles and calculates a credibility score.

[2109] Specific actions

[2110] The generative AI model analyzes the text and metadata of the input news article and calculates a credibility score.

[2111] input

[2112] Parse prompts, news article body text and metadata.

[2113] output

[2114] Reliability scores and analysis results.

[2115] Step 6:

[2116] The server transmits the calculated reliability score and the analysis result to the user terminal.

[2117] Specific actions

[2118] The server returns the reliability score and analysis results in JSON format to the user device via the REST API set at the endpoint.

[2119] input

[2120] Reliability scores and analysis results.

[2121] output

[2122] The reliability score and analysis results sent to the user's device.

[2123] Step 7:

[2124] Based on the information received by the terminal, a warning message is displayed on the user interface.

[2125] Specific actions

[2126] The user interface (HTML, CSS, JavaScript) receives the reliability score and analysis results from the server and displays a warning message such as "This news article is not reliable" as a popup or banner.

[2127] input

[2128] Reliability scores and analysis results.

[2129] output

[2130] The warning message displayed in the user interface.

[2131] Detecting edited video and image data

[2132] Step 1:

[2133] A user uploads a video or image file onto the Internet.

[2134] Specific actions

[2135] The user interface displays a file selection dialog and the user selects a file.

[2136] input

[2137] Video or image files selected by the user.

[2138] output

[2139] The selected file will be uploaded to the system.

[2140] Step 2:

[2141] The terminal transmits the selected video or image data to the server.

[2142] Specific actions

[2143] Use the JavaScript FormData object to send the selected file to the server via a POST request.

[2144] input

[2145] Video or image files.

[2146] output

[2147] Video or image files sent to the server.

[2148] Step 3:

[2149] The server sends the received video or image data to the generative AI model, which generates an analysis prompt.

[2150] Specific actions

[2151] The server inputs the selected file data into a generation AI (e.g., DALL-E) and generates a prompt saying, "Please detect any editing traces in this image."

[2152] input

[2153] Uploaded video or image files.

[2154] output

[2155] The generated analysis prompt.

[2156] Step 4:

[2157] The generative AI model checks the video or image data for signs of editing and sends the results back to the server.

[2158] Specific actions

[2159] The AI ​​detects unnatural changes, color inconsistencies, and shadow discontinuities in the image and sends the analysis results back to the server.

[2160] input

[2161] Video or image file, analysis prompt.

[2162] output

[2163] Analysis results including traces of editing.

[2164] Step 5:

[2165] The server transmits the analysis results to the user terminal.

[2166] Specific actions

[2167] The server returns the analysis results in JSON format to the user device via the REST API set at the endpoint.

[2168] input

[2169] Analysis results including traces of editing.

[2170] output

[2171] Analysis results sent to the user's device.

[2172] Step 6:

[2173] Based on the analysis results, the device displays a warning message on the user interface.

[2174] Specific actions

[2175] The user interface receives the analysis results and displays a warning message such as "This footage may be faked" as a pop-up or banner.

[2176] input

[2177] Analysis results including traces of editing.

[2178] output

[2179] The warning message displayed in the user interface.

[2180] Social Media Monitoring

[2181] Step 1:

[2182] The browser extension monitors your online activity as you browse social media sites.

[2183] Specific actions

[2184] The browser extension collects the URLs of the sites you visit and the scripts those sites run.

[2185] input

[2186] Information about the social media sites you visit.

[2187] output

[2188] Site Information Collected.

[2189] Step 2:

[2190] The terminal sends the collected site information to the server.

[2191] Specific actions

[2192] The browser extension sends the collected data in JSON format to the server via a POST request.

[2193] input

[2194] Site Information Collected.

[2195] output

[2196] Site information sent to the server.

[2197] Step 3:

[2198] The server analyzes the site information it receives to determine whether unauthorized data collection or tracking is occurring.

[2199] Specific actions

[2200] The server analyzes the collected data against industry-standard security rules (for example, the OWASP threat list).

[2201] input

[2202] Received site information.

[2203] output

[2204] Verification of fraudulent data collection or tracking.

[2205] Step 4:

[2206] The server transmits the determination result to the user terminal.

[2207] Specific actions

[2208] The server returns the result of the judgment in JSON format to the user device via the REST API set at the endpoint.

[2209] input

[2210] Verification of fraudulent data collection or tracking.

[2211] output

[2212] The judgment result sent to the user terminal.

[2213] Step 5:

[2214] Based on the judgment result, the terminal displays a warning message on the user interface.

[2215] Specific actions

[2216] The user interface receives the judgment result and displays a warning message such as "This site is collecting data illegally" as a pop-up or banner.

[2217] input

[2218] Judgment result.

[2219] output

[2220] The warning message displayed in the user interface.

[2221] (Application example 1)

[2222] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2223] The decline in the reliability of information on the Internet and the resulting spread of misinformation and fake news have become a social problem. Furthermore, the increasing number of edited videos and images, as well as the increasing number of illegal data collection and tracking practices on social media, are making it difficult for users to determine the authenticity of information. These problems must be resolved quickly and accurately, and users must be provided with reliable information.

[2224] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[2225] In this invention, the server includes means for collecting data via a network and acquiring news article text and metadata, means for analyzing the acquired text and metadata and calculating a reliability score for the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for receiving and analyzing video or image data and detecting editing and unauthorized manipulation, and means for monitoring and determining unauthorized data collection and tracking on social media sites and displaying a warning to the user. This allows users to quickly evaluate the reliability of information on the Internet and be protected from unauthorized data collection.

[2226] A "network" is a collection of computers and communication devices that allow information and data to be sent and received.

[2227] "Data" refers to the information being analyzed, such as news articles, videos, images, and social media information.

[2228] A "news article" is written information such as a news report posted on the Internet.

[2229] "Body" refers to the main content of a news article.

[2230] "Metadata" is supplemental information and attribute data associated with a news article.

[2231] A "trust score" is a numerical assessment of the trustworthiness of a particular article or piece of information.

[2232] A "warning" is a message that is displayed to the user to call attention to the situation.

[2233] "Video" is data that contains moving visual information.

[2234] An "image" is data containing still visual information.

[2235] "Editing evidence" refers to evidence of changes that indicate an image or video has been modified.

[2236] "Social media" is an online platform for sharing information between users.

[2237] "Unauthorized data collection" refers to the illegal act of collecting and using user information without permission.

[2238] "Tracking" is the act of following a user's online activities.

[2239] A "user terminal" is a device used to view information and receive alerts.

[2240] A "generative AI model" is a machine learning model for performing natural language processing and data analysis.

[2241] A "prompt sentence" is a sentence used to input instructions or questions into a generative AI model.

[2242] This invention provides a system for evaluating the reliability of information on the Internet. The system analyzes news articles, video and image data, and information on social media, and conveys the reliability to users.

[2243] Program Generation

[2244] The following describes an outline of a program and its processing for realizing the present invention.

[2245] Hardware / Software used

[2246] Hardware: User's smartphone, server

[2247] Software: Mobile apps, browser extensions, generative AI models (e.g., GPT-4), data analysis tools

[2248] News article credibility assessment

[2249] When a user views a news article on their smartphone, the URL of that news article is automatically sent to a server. A generative AI model (e.g., GPT-4) on the server retrieves and analyzes the news article text and metadata from the URL. Based on the analysis results, a reliability score is calculated and sent to the user's smartphone. If the reliability score is low, the user's smartphone displays a warning message such as "This news article is unreliable."

[2250] Examples:

[2251] When a user clicks on a news article shared on a social networking site, the URL is sent to a server, which analyzes the news article and, if the reliability score is low, displays a warning message on the user's smartphone saying, "This news article is unreliable."

[2252] Prompt for the generative AI model:

[2253] News article URL: [example.com / news123]

[2254] Rate the article's credibility and calculate a credibility score, and explain why it contains exaggerated or unreliable content.

[2255] Detecting traces of editing of video and image data

[2256] When a user views a video or image file on their smartphone, the file is automatically uploaded to a server. A generative AI model on the server analyzes the video or image and detects any signs of editing. If there is a problem with its authenticity, that information is sent to the user's smartphone, warning them that "this video may be faked."

[2257] Examples:

[2258] When a user downloads a video shared on a social networking site and uploads it to the system, the server analyzes the video, and if any fraud is detected, a warning message will appear on the smartphone saying, "This video may be faked."

[2259] Social Media Monitoring

[2260] When a user visits a social media site in their browser, the browser extension monitors the user's online activity and sends the site information to a server. An analysis tool on the server identifies any unauthorized data collection or tracking attempts and sends the results to the user's smartphone. If any unauthorized activity is detected, a warning message stating "This site is engaging in unauthorized data collection" is displayed.

[2261] Examples:

[2262] When a user visits a social media site, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, the user's smartphone will display a warning message stating, "This site is engaging in unauthorized data collection."

[2263] In this way, users can quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection, enabling them to enjoy a safer and more reliable information environment.

[2264] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[2265] Program flow:

[2266] News article credibility assessment

[2267] Step 1:

[2268] The user terminal inputs the URL of the news article.

[2269] The entered URL is sent from the user's smartphone to the server.

[2270] Step 2:

[2271] The server retrieves the text and metadata of the news article based on the received URL.

[2272] The server uses a web scraping tool to retrieve data from the URL.

[2273] Step 3:

[2274] The server parses the retrieved text and metadata.

[2275] A generative AI model (e.g., GPT-4) is used to calculate a confidence score for the text.

[2276] Step 4:

[2277] The server transmits the calculated reliability score to the user terminal.

[2278] The reliability score and analysis results are displayed on the user's smartphone.

[2279] Step 5:

[2280] If the user terminal has a low reliability score, a warning message is displayed.

[2281] Users may see a message that reads, "This news article is not reliable."

[2282] Detecting traces of editing of video and image data

[2283] Step 1:

[2284] The user uploads a video or image to the device.

[2285] The uploaded data is automatically sent to the server.

[2286] Step 2:

[2287] The server analyzes the received video or image data.

[2288] Generative AI models are used to detect edits (e.g., unnatural color tones and shadows).

[2289] Step 3:

[2290] The server transmits the analysis results to the user terminal.

[2291] A warning message appears on the user's smartphone saying, "This video may be faked."

[2292] Step 4:

[2293] The user terminal displays a warning message and presents the analysis results.

[2294] Users can see the specific locations and reasons for unauthorized edits.

[2295] Social Media Monitoring

[2296] Step 1:

[2297] A user browses a social media site.

[2298] Browser extensions monitor your online activity and collect site information.

[2299] Step 2:

[2300] The server analyzes the collected site information.

[2301] Generative AI models identify potential fraudulent data collection and tracking.

[2302] Step 3:

[2303] The server transmits the determination result to the user terminal.

[2304] A warning message appears on the user's smartphone saying, "This site is collecting data illegally."

[2305] Step 4:

[2306] The user terminal displays a warning message and provides details of the fraudulent activity.

[2307] Users will be given a detailed explanation of what fraudulent activity is taking place.

[2308] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[2309] ---

[2310] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[2311] Evaluating the credibility of news articles and using sentiment engines

[2312] When a user views a news article online, the user's device sends the news article's URL to a server. The server retrieves the news article's text and metadata from the received URL. A generation AI located on the server analyzes the news article and calculates a reliability score based on the text content, source reliability, past history, etc. The user's emotion engine then recognizes the user's emotional state (excited, angry, calm, etc.). The calculated reliability score and analysis results are sent to the user's device, and the way the warning is displayed is adjusted based on the user's emotional state detected by the emotion engine. When a warning message such as "This news article is unreliable" is displayed, if the user is angry, the warning will be displayed more gently, and if the user is calm, a more direct warning will be displayed.

[2313] Specific examples

[2314] When a user clicks on a news article shared on a social networking site, the URL is sent to the server. The server analyzes the news article, and if the reliability score is low, a warning message is displayed on the user's device. At the same time, the emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a gentle warning if the user is excited.

[2315] Detecting edited video and image data and using emotion engines

[2316] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generation AI checks the video or image for signs of editing and identifies any unauthorized changes. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. The warning, "This video may be faked," is adjusted according to the user's state.

[2317] Specific examples

[2318] Users download videos posted on social media and upload them to the system. The server analyzes the video to determine whether there are any signs of unauthorized editing. If unauthorized editing is detected, the emotion engine recognizes the user's emotions and displays a gentle warning message if the user is feeling angry.

[2319] Social media monitoring and sentiment engine utilization

[2320] When a user visits a social media site in a browser, a browser extension is used to monitor the user's online activity. The user's device sends information about the site being accessed to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. The result of the determination is sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message is displayed encouraging them to stay calm. The warning, "This site is engaging in unauthorized data collection," is adjusted according to the user's state.

[2321] Specific examples

[2322] As a user browses social media sites, the browser extension monitors the site's scripts and requests. If the server determines that the site is engaging in unauthorized data collection, an emotion engine recognizes the user's emotions and displays a direct warning if the user is calm, or a more gentle message if the user is impatient.

[2323] In this way, users can quickly assess the reliability of information provided on the Internet and are protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into account, providing more effective and responsive warnings. This allows users to enjoy a safer and more reliable information environment.

[2324] ---

[2325] This concludes the description of the embodiment of the present invention. This system can more effectively ensure the accuracy of information and provide appropriate warnings to users by evaluating the reliability of news articles, detecting edits in video and image data, and monitoring social media, as well as taking into account the emotional state of the user.

[2326] The processing flow will be explained below.

[2327] Evaluating the credibility of news articles and using sentiment engines

[2328] Step 1:

[2329] A user clicks on a news article link, and the user's device sends the news article URL to the server.

[2330] Step 2:

[2331] The server retrieves the HTML data of the news article from the received URL.

[2332] Step 3:

[2333] The server extracts the news article text and metadata from the HTML data it has obtained.

[2334] Step 4:

[2335] The generation AI placed on the server analyzes the text of news articles using natural language processing (NLP) technology and evaluates the content of the article, the reliability of the source, past history, etc.

[2336] Step 5:

[2337] The server calculates a reliability score based on the analysis results.

[2338] Step 6:

[2339] The server queries the user's emotion engine for the user's most recent emotional state.

[2340] Step 7:

[2341] The user's device activates an emotion engine to detect the user's emotional state (excitement, anger, calmness, etc.).

[2342] Step 8:

[2343] The emotion engine transmits the user's emotional state to the server.

[2344] Step 9:

[2345] The server adjusts how the warning message is displayed depending on the confidence score and the detected emotional state.

[2346] Step 10:

[2347] The server sends a reliability score and a warning message to the user's device.

[2348] Step 11:

[2349] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[2350] Detecting edited video and image data and using emotion engines

[2351] Step 1:

[2352] A user uploads video or image data to the system.

[2353] Step 2:

[2354] The user's terminal transmits the uploaded file data to the server.

[2355] Step 3:

[2356] The server receives the file data.

[2357] Step 4:

[2358] The server's generated AI analyzes video and image data and identifies traces of unauthorized editing.

[2359] Step 5:

[2360] The server creates a warning message based on the analysis results.

[2361] Step 6:

[2362] The server queries the user's emotion engine for the emotional state.

[2363] Step 7:

[2364] The user's terminal activates an emotion engine to detect the user's emotional state.

[2365] Step 8:

[2366] The emotion engine transmits the user's emotional state to the server.

[2367] Step 9:

[2368] The server tailors the warning message based on the analysis results and emotional state.

[2369] Step 10:

[2370] The server sends the analysis results and warning messages to the user's terminal.

[2371] Step 11:

[2372] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[2373] Social media monitoring and sentiment engine utilization

[2374] Step 1:

[2375] A user browses to a social media site.

[2376] Step 2:

[2377] A browser extension is launched on the user's device and collects information about the sites they are viewing.

[2378] Step 3:

[2379] The user's terminal transmits the collected site information to the server.

[2380] Step 4:

[2381] The server analyzes the received site information.

[2382] Step 5:

[2383] The server inspects your site's scripts and requests to determine if there are any unauthorized attempts at data collection or tracking.

[2384] Step 6:

[2385] The server generates a verdict and a warning message.

[2386] Step 7:

[2387] The server queries the user's emotion engine for the emotional state.

[2388] Step 8:

[2389] The user's terminal activates an emotion engine to detect the user's emotional state.

[2390] Step 9:

[2391] The emotion engine transmits the user's emotional state to the server.

[2392] Step 10:

[2393] The server adjusts the warning message based on the result of the assessment and the emotional state.

[2394] Step 11:

[2395] The server sends the judgment result and a warning message to the user's terminal.

[2396] Step 12:

[2397] The user's terminal displays a warning message and conveys the warning in an appropriate format based on the user's emotional state.

[2398] ---

[2399] In this way, by taking into account the user's emotional state in addition to assessing the reliability of news articles, detecting traces of editing in video and image data, and monitoring social media, we can more effectively ensure the accuracy of information and provide appropriate warnings to users.

[2400] Example 2

[2401] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2402] The internet is flooded with information, often containing unreliable news articles and fraudulently manipulated video and images. Social media sites also sometimes conduct fraudulent data collection and tracking. This information not only causes confusion and anxiety among users, but can also lead to social unrest. Therefore, it is important to evaluate the reliability of this information and provide appropriate warnings to users. However, conventional systems often display uniform warnings without considering the user's emotional state, which can detract from the user experience.

[2403] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[2404] In this invention, the server includes means for acquiring news article text and metadata, means for analyzing the acquired text and metadata and using a generative AI model to calculate a reliability score for the information, means for using an emotion engine to recognize the user's emotional state, means for adjusting a warning based on the calculated reliability score and the user's emotional state and sending the adjusted warning message to the user terminal, and means for displaying the adjusted warning message on the user terminal. This makes it possible to evaluate the reliability of news articles, video / image data, and social media site information, and to provide flexible warnings that take the user's emotional state into consideration.

[2405] "Means for obtaining news article text and metadata" refers to a system for collecting the text of news articles published on the Internet and related metadata such as publication date, author information, and URL.

[2406] "Means for using generative AI models" refers to a mechanism for using generative artificial intelligence models to analyze collected data and output the results.

[2407] An "emotion engine for recognizing the user's emotional state" is an engine that can identify and analyze the user's emotions, and is a mechanism for determining whether the user is excited, angry, calm, etc.

[2408] The "means for calculating the reliability score of information" is a system that analyzes the content and metadata of news articles, video and image data, and expresses the reliability of that information as a quantified score.

[2409] The "means for adjusting and sending a warning to a user terminal" is a mechanism for adjusting the content and display method of a warning message based on the calculated reliability score and the recognized emotional state of the user, and sending it to the user's terminal.

[2410] The "means for displaying the adjusted warning message on the user terminal" is a mechanism for displaying the adjusted warning message sent from the server on the user terminal.

[2411] "Means for receiving video or image data" refers to a mechanism for receiving video or image data uploaded by a user.

[2412] "Means for identifying traces of unauthorized editing" refers to a mechanism for analyzing received video or image data and checking for traces of editing or manipulation.

[2413] "Means for collecting site information" refers to a mechanism for collecting information on the social media sites that users access.

[2414] "Means for determining attempts at unauthorized data collection or tracking" refers to a mechanism for analyzing collected site information to determine whether the site is attempting unauthorized data collection or tracking of users.

[2415] This invention relates to a system that ensures the accuracy of information on the Internet and provides warnings taking into account the user's emotional state. The system aims to ensure the reliability and accuracy of information by evaluating and analyzing news articles, video and image data, and information from social media, and by recognizing the user's emotions.

[2416] Evaluating the credibility of news articles and using sentiment engines

[2417] When a user browses a news article online, the user's device sends the URL of the news article to a server. The server retrieves the news article's text and metadata from the URL. A generative AI deployed on the server analyzes the news article and calculates a credibility score based on the text's content, the credibility of the source, and past history. Specifically, a generative AI model (such as GPT-4 for natural language processing) is used to analyze the text and assign a credibility score. An emotion engine (such as IBM Watson's Tone Analyzer) then recognizes the user's emotional state. The calculated credibility score and analysis results are sent to the user's device, and the display of the warning is adjusted based on the user's emotional state detected by the emotion engine. For example, when a warning message stating "This news article is unreliable" is displayed to a user, if the user is angry, the warning will be displayed more gently, whereas if the user is calm, a more direct warning will be displayed.

[2418] Example prompt sentence:

[2419] "Please parse the following news article URL and calculate a credibility score: [news article URL]"

[2420] Detecting edited video and image data and using emotion engines

[2421] When a user views video or image files on the Internet, they upload these files to the system. The user's device sends the video or image data to the server, which analyzes it. The server's generated AI checks the video or image for signs of editing and identifies any unauthorized changes. For example, it uses Adobe Photoshop's image analysis function to detect editing. The analysis results are sent back to the user's device, and an emotion engine recognizes the user's emotional state. If the user shows signs of heightened emotion, a gentle warning is displayed. For example, the warning "This video may be faked" is adjusted according to the user's state.

[2422] Example prompt sentence:

[2423] "Analyze this video file to check for any signs of unauthorized editing: [video file path]"

[2424] Social media monitoring and sentiment engine utilization

[2425] When a user browses social media sites in a browser, their online activity is monitored using a browser extension. The user's device sends information about the sites they are visiting to a server. The server analyzes the received site information and determines whether the site is engaging in unauthorized data collection or tracking. For example, this uses the monitoring function of Ghostery. The results of this determination are sent to the user's device, and an emotion engine recognizes the user's emotional state. If the user is feeling anxious, a warning message urging them to stay calm is displayed. For example, the warning "This site is engaging in unauthorized data collection" is adjusted according to the user's state.

[2426] Example prompt sentence:

[2427] "Check if this social media site is illegally collecting data: [site URL]"

[2428] The goal of this system is to evaluate the reliability of news articles, video and image data, and social media information, helping users easily determine the accuracy of the information, while also taking into account the user's emotional state to provide more effective warnings.

[2429] The flow of the identification process in the second embodiment will be described with reference to FIG.

[2430] Evaluating the credibility of news articles and using sentiment engines

[2431] Step 1:

[2432] A user clicks on a news article URL.

[2433] (Input) The URL the user clicks on.

[2434] (Output) URL information clicked on the user's device.

[2435] Users click on the URL of the news article they want to view in their web browser.

[2436] Step 2:

[2437] The device sends the URL information to the server.

[2438] (Input) The URL of the news article that was clicked.

[2439] (Output) The URL information sent to the server.

[2440] The user's terminal sends the URL of the clicked news article to the server as an HTTP request.

[2441] Step 3:

[2442] The server retrieves the news article text and metadata.

[2443] (Input) The URL of the received news article.

[2444] (Output) The news article text and metadata.

[2445] The server uses the received URL to obtain the text of the news article and metadata (publication date, author name, etc.) from the news site using scraping or an API.

[2446] Step 4:

[2447] The server instructs the generated AI to analyze.

[2448] (Input) News article text and metadata.

[2449] (Output) Analysis instructions to the generating AI.

[2450] The server sends the text and metadata of the retrieved news article as input to the generation AI, instructing it to analyze it.

[2451] Step 5:

[2452] The generative AI calculates a reliability score.

[2453] (Input) News article text and metadata.

[2454] (Output) The confidence score.

[2455] The generative AI (e.g., GPT-4) analyzes the text and metadata of the received news article and calculates a credibility score based on the content of the text, the reliability of the source, past history, etc.

[2456] Step 6:

[2457] An emotion engine recognizes the user's emotional state.

[2458] (Input) User behavior data (viewing time, scrolling speed, etc.).

[2459] (Output) The user's emotional state.

[2460] An emotion engine (e.g., IBM Watson Tone Analyzer) analyzes the user's behavioral data and recognizes the user's current emotional state (excited, angry, calm, etc.).

[2461] Step 7:

[2462] The server sends the analysis results and warning messages to the terminal.

[2463] (Input) Confidence score and user's emotional state.

[2464] (Output) The warning message.

[2465] The server adjusts the content of the warning message based on the reliability score calculated by the generation AI and the user's emotional state recognized by the emotion engine, and sends it to the user's device.

[2466] Step 8:

[2467] The terminal displays a warning message to the user.

[2468] (Input) The warning message received from the server.

[2469] (Output) The warning message to be displayed.

[2470] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the news article.

[2471] Detecting edited video and image data and using emotion engines

[2472] Step 1:

[2473] A user uploads video or image data to the system.

[2474] (Input) Video or image file.

[2475] (Output) The data uploaded to the system.

[2476] Users upload video or image files to the system via the Internet.

[2477] Step 2:

[2478] The device sends the data to the server.

[2479] (Input) Uploaded video or image data.

[2480] (Output) The data sent to the server.

[2481] The user's terminal transmits the uploaded video or image data to the server.

[2482] Step 3:

[2483] The server instructs the generated AI to analyze.

[2484] (Input) Received video or image data.

[2485] (Output) Analysis instructions to the generating AI.

[2486] The server sends the received data to the generation AI and instructs it to analyze it.

[2487] Step 4:

[2488] Generative AI checks for signs of editing in videos and images.

[2489] (Input) Video or image data.

[2490] (Output) Analysis results for edit traces.

[2491] The generating AI analyzes the received video and image data and checks for signs of editing or unauthorized processing.

[2492] Step 5:

[2493] An emotion engine recognizes the user's emotional state.

[2494] (Input) User behavior data (viewing time, reaction speed, etc.).

[2495] (Output) The user's emotional state.

[2496] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[2497] Step 6:

[2498] The server sends the analysis results and warning messages to the terminal.

[2499] (Input) Analysis results of editing traces and the user's emotional state.

[2500] (Output) The warning message.

[2501] The server adjusts the content of the warning message based on the analysis results and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[2502] Step 7:

[2503] The terminal displays a warning message to the user.

[2504] (Input) The warning message received from the server.

[2505] (Output) The warning message to be displayed.

[2506] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning regarding the reliability of the video or image.

[2507] Social media monitoring and sentiment engine utilization

[2508] Step 1:

[2509] A user browses a social media site.

[2510] (Input) Social media site URL.

[2511] (Output) Site browsing information on the user's device.

[2512] Users use a web browser to browse social media sites.

[2513] Step 2:

[2514] The site information that the terminal is accessing is sent to the server.

[2515] (Input) Information about the site you are viewing.

[2516] (Output) Site information sent to the server.

[2517] The browser extension collects information about the social media sites you visit and sends it to a server.

[2518] Step 3:

[2519] The server analyzes the site information.

[2520] (Input) Received site information.

[2521] (Output) Site analysis results.

[2522] The server analyzes the received information from the social media site and evaluates its content.

[2523] Step 4:

[2524] The server determines whether or not unauthorized data collection has occurred.

[2525] (Input) Site analysis results.

[2526] (Output) The result of the invalid data collection determination.

[2527] The server uses the site information to determine whether unauthorized data collection or tracking is occurring.

[2528] Step 5:

[2529] An emotion engine recognizes the user's emotional state.

[2530] (Input) User behavior data (viewing time, reaction speed, etc.).

[2531] (Output) The user's emotional state.

[2532] The emotion engine analyzes the user's behavioral data and recognizes the emotional state the user is currently experiencing.

[2533] Step 6:

[2534] The server sends the judgment result and a warning message to the terminal.

[2535] (Input) The result of the fraudulent data collection and the user's emotional state.

[2536] (Output) The warning message.

[2537] The server adjusts the content of the warning message based on the judgment result and the user's emotional state recognized by the emotion engine, and sends it to the user's terminal.

[2538] Step 7:

[2539] The terminal displays a warning message to the user.

[2540] (Input) The warning message received from the server.

[2541] (Output) The warning message to be displayed.

[2542] The user's terminal displays the warning message sent from the server on the screen, providing the user with a warning about the trustworthiness of the social media site.

[2543] (Application example 2)

[2544] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2545] In today's Internet environment, there is a lot of false information and incorrect data, making it difficult for users to determine whether or not to trust this information. Displaying appropriate warning messages based on the user's emotional state is also problematic. Conventional systems display uniform warning messages that ignore the user's emotional state, which can prevent users from receiving warnings appropriately.

[2546] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for acquiring the text and metadata of news articles, means for analyzing the acquired text and metadata and calculating a reliability score of the information, means for transmitting the calculated reliability score to a user terminal and displaying a warning, means for recognizing the user's emotional state and adjusting the display of the warning message based on the emotional state, and means for analyzing the URL and metadata of video content, calculating a reliability score, and displaying a warning message. This allows users to quickly evaluate the reliability of information provided on the Internet and be protected from unauthorized data collection. Furthermore, the emotion engine takes the user's emotional state into consideration, providing more effective and responsive warnings.

[2547] A "network" is an infrastructure that allows multiple computers and devices to communicate with each other.

[2548] "Data" is a coded set of letters and numbers that represent specific information.

[2549] A "news article" is a piece of writing that reports social, economic, political, or cultural events or information.

[2550] "Body" is the text portion that constitutes the main content of a news article.

[2551] "Metadata" is data that describes information and attributes about data.

[2552] A "trustworthiness score" is an evaluation index that quantifies the accuracy and reliability of information.

[2553] A "user terminal" is a device used by a user to obtain and operate information.

[2554] A "warning" is a message intended to inform the user of a potential risk or problem.

[2555] "Emotional state" refers to the user's current psychological response or emotional state.

[2556] A "warning message" is textual or graphical information displayed to inform the user of a warning.

[2557] "Video content" refers to media data that is viewed as moving images.

[2558] A "URL" is an address that identifies the location of a web page or other Internet resource.

[2559] "Analysis" is the process of examining data in detail to reveal its structure and meaning.

[2560] "Unauthorized editing" is any deliberate alteration made to a footage or image that is false or misleading.

[2561] "Online activities" refer to the various operations and actions that users perform using the Internet.

[2562] "Data collection" is the act of collecting and storing information for a specific purpose.

[2563] "Tracking" refers to any technology or process that follows or records a user's online actions or movements.

[2564] A "generative AI model" is a model that is trained to solve a specific problem using artificial intelligence techniques.

[2565] A "prompt" is a text instruction entered into a generative AI model.

[2566] This invention relates to a system that evaluates the reliability of information in content distribution services and displays appropriate warnings according to the user's emotional state. This system provides users with a safe and reliable information environment by monitoring news articles, video content, video and image data, and social media sites.

[2567] Key Components

[2568] The main components of this system are as follows:

[2569] 1. Data collection method: A module for collecting URLs and metadata of news articles and video content. This includes a mechanism for sending data from the user's device to the server. 【257...

Claims

1. To collect data over the network and evaluate the reliability of news articles, a means for obtaining news article text and metadata; means for analyzing the retrieved text and metadata and calculating a reliability score for the information; The system includes a means for transmitting the calculated reliability score to a user terminal and displaying a warning.

2. Analyze video or images to detect editing and unauthorized manipulation; means for receiving uploaded video or image data; means for analyzing the received data and identifying traces of unauthorized editing; 2. The system according to claim 1, further comprising means for transmitting the analysis results to a user terminal and displaying a warning.

3. To monitor social media sites to detect unauthorized data collection and tracking; means for monitoring users' online activities and collecting site information; A means for analyzing collected site information to determine unauthorized data collection or tracking attempts; and 2. The system according to claim 1, further comprising means for transmitting the determination result to a user terminal and displaying a warning.

Images