System

Abstract

A system is provided.SOLUTION: A system for determining whether content is created by a production AI, comprising: means for receiving content from a user; means for preprocessing the received content to extract textual data; means for extracting features from the preprocessed textual data; means for inputting the extracted features into a production AI determination algorithm to obtain a determination result; and means for notifying the user of the determination result.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] Among the vast amount of information spreading across the internet, an increasing amount of content is being created by generative AI. This AI-generated content can become a breeding ground for false information and fake news, posing a problem of undermining the reliability of information. In particular, in industries where information reliability is important, such as media companies and advertising agencies, there is a need to distinguish between content created by generative AI and content created by humans. Therefore, the objective of this invention is to accurately identify content created by generative AI and improve the reliability of information. [Means for solving the problem]

[0005] The present invention provides a system for determining whether content was created by a generative AI. This system includes a means for receiving content from a user, a means for preprocessing the received content to extract text data, a means for extracting features from the preprocessed text data, a means for inputting the extracted features into a generative AI determination algorithm to obtain a determination result, and a means for notifying the user of the determination result, thereby enabling highly accurate determination of whether content was created by a generative AI. Furthermore, by including a means for extracting metadata and a means for generating a report based on the determination result, the system provides more reliable information.

[0006] "Receiving" means taking in data or content sent by a user.

[0007] "Preprocessing" is the preparatory work of analyzing received data or content and extracting necessary information.

[0008] "Text data" refers to sentences and character-based information extracted through preprocessing.

[0009] "Features" are patterns and attributes extracted from text data that are important in data analysis.

[0010] A "generative AI determination algorithm" is a machine learning model or mathematical method for identifying whether given data was created by generative AI.

[0011] "Determination result" is a conclusion reached using the AI ​​generation determination algorithm, and indicates whether or not the content was generated by AI.

[0012] "Notification" refers to the act of informing the user of the judgment result, and includes means such as displaying it on the dashboard or sending an email.

[0013] "Metadata" is supplementary information about content, and includes attributes such as creator information and posting date and time.

[0014] A "report" is a document or data summarizing the judgment results and their grounds, and is the final notification to the user. [Brief explanation of the drawings]

[0015] [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

[0016] 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.

[0017] First, the terms used in the following description will be explained.

[0018] 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).

[0019] 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.

[0020] 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.

[0021] 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.

[0022] 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."

[0023] [First embodiment]

[0024] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0025] 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.

[0026] 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).

[0027] 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.

[0028] 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.

[0029] 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.

[0030] 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.

[0031] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] 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.

[0033] 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.

[0034] 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.

[0035] 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."

[0036] The present invention relates to a system that accurately judges content created by generative AI, and specific embodiments thereof will be described below with specific examples.

[0037] System configuration

[0038] This system focuses on receiving content sent by users, preprocessing it to extract text data, inputting features from the text data into a generative AI judgment algorithm, and determining whether the content was created by generative AI.

[0039] Program processing

[0040] Receiving content

[0041] When a user sends content such as an article, advertisement, or review from a terminal, the server receives the content. For example, when a reporter from a media company posts a new article to the system, the article data is sent to the server.

[0042] Data Preprocessing

[0043] The server preprocesses the received content and extracts text data. Preprocessing includes text cleaning and parsing. If the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) to extract text from them.

[0044] Feature extraction

[0045] From the preprocessed text data, the server extracts features, which include word frequency, sentence structure, specific patterns of writing style, etc. Specific features are selected based on a machine learning model.

[0046] Execution of generative AI decision algorithm

[0047] The server inputs the extracted features into a generative AI determination algorithm, which uses a pre-trained model to determine whether the content was created by generative AI. For example, the server uses a machine learning model such as a multilayer neural network or a decision tree to make the determination.

[0048] Output of judgment results and report generation

[0049] The server generates a report based on the results of its assessment. This report includes whether the content was created by a generative AI and the reasons for this. For example, if it concludes that "this article was created by a generative AI," it will cite specific stylistic patterns and unnatural phrasing as reasons for this.

[0050] Notification of results

[0051] The server notifies the user of the generated report. Possible notification methods include sending an email or displaying it on a dashboard. For example, a reporter from a media company can check the dashboard to confirm that the article he or she posted was not generated by a generative AI.

[0052] Specific examples

[0053] Media company use cases

[0054] A reporter (user) submits a new article to the system from their device. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[0055] Example of use at an advertising agency

[0056] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or other means and can confirm them.

[0057] The above is a specific embodiment for carrying out the present invention. By using this system, content can be identified by the generation AI with high accuracy, and the reliability of the information can be ensured.

[0058] The processing flow will be explained below.

[0059] Step 1:

[0060] Users upload content from their devices. For example, journalists post articles and advertising creators upload advertising content.

[0061] Step 2:

[0062] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[0063] Step 3:

[0064] The content received by the server is preprocessed. This preprocessing involves text extraction. If the content is in a specific format like HTML or PDF, only the text data is extracted and cleaned up. Unnecessary tags and special characters are removed.

[0065] Step 4:

[0066] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[0067] Step 5:

[0068] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[0069] Step 6:

[0070] The server generates features from the extracted text data, including word frequency, sentence structure, and specific phrasing patterns. Stylistic patterns and unnatural expressions in the text are also analyzed here.

[0071] Step 7:

[0072] The server inputs the generated features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI.

[0073] Step 8:

[0074] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was created by generative AI, including information on which features were important.

[0075] Step 9:

[0076] The server generates a report based on the findings, including the findings, rationale, and optional metadata, such as a conclusion like "This article was not created by a generative AI."

[0077] Step 10:

[0078] The server notifies the user of the generated report, which can be done by email or displayed on a dashboard. The user can then access the dashboard to check the results.

[0079] These are the specific processing steps of the content assessment system using generative AI. At each step, the server, device, and user play their respective roles, and the system as a whole functions to increase the reliability of information.

[0080] Example 1

[0081] 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."

[0082] Conventional systems have low accuracy in determining content created by generative AI, making it difficult to obtain reliable results. Furthermore, when content is provided in a variety of formats (text, images, video), there is a lack of a way to consistently analyze them, making it difficult to make a comprehensive determination. This has made it difficult to ensure the reliability of information.

[0083] 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.

[0084] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for generating a report based on the judgment result, means for notifying the user of the judgment result, and means for using character recognition technology to extract text from images or videos. This enables highly accurate analysis of various types of content and highly reliable content judgment by the generation AI.

[0085] "Content" refers to information such as articles, advertisements, reviews, etc. submitted by users, and may include various formats such as text, images, and video.

[0086] A "user" is a person or organization that uses this system to send content and receives the results of the evaluation.

[0087] A "server" is a computer system that receives content and performs processes such as preprocessing, feature extraction, execution of a determination algorithm, report generation, and notification of results.

[0088] "Preprocessing" refers to the process of extracting text data from received content and performing cleaning and analysis.

[0089] "Text Data" means textual information extracted from Content, including text with HTML tags removed and text extracted from images or video using OCR technology.

[0090] "Features" are data characteristics extracted from preprocessed text data, and include word frequency, sentence structure, and specific patterns of writing style.

[0091] A "generative AI determination algorithm" is an algorithm that uses a pre-trained model to determine whether content was created by generative AI.

[0092] A "report" is a document that describes the judgment results and their rationale, and is generated in HTML or PDF format.

[0093] "Optical Character Recognition (OCR)" is a technology that extracts character data from images and videos.

[0094] "Judgment result" is the output of the content judgment obtained by the generative AI judgment algorithm.

[0095] "Notification" refers to informing the user of the judgment result, and includes methods such as sending an email or displaying it on a dashboard.

[0096] The present invention relates to a system that judges content created by generative AI with high accuracy, and will be explained below with specific examples.

[0097] Receiving content

[0098] When a user sends content such as an article, advertisement, or review from their device, the server receives the content. For example, when a journalist from a media company posts a new article to the system, the article data is sent to the server. The server authenticates the user and verifies the source of the content. Specifically, it parses the JSON-formatted data packet and extracts the article ID and poster ID.

[0099] Data Preprocessing

[0100] The server preprocesses the received content and extracts text data. Preprocessing includes cleaning and parsing the text. For example, if the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) technology to extract text. Specifically, it uses the Python Tesseract library to extract characters from images.

[0101] Feature extraction

[0102] The server extracts features from the preprocessed text data. These include word frequency, sentence structure, and stylistic patterns. The server uses the NLTK library to count frequently occurring words in the text and encode the words (Word2Vec). It also performs POS tagging and dependency structure analysis to detect stylistic patterns. Specifically, it uses the SpaCy library to perform morphological analysis of sentences and extract important grammatical structures as features.

[0103] Execution of generative AI decision algorithm

[0104] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained model to determine whether the content was created by generative AI. For example, it inputs the features and performs classification using a multi-layer neural network (using TensorFlow or PyTorch). This classification model is previously trained on data from generative AI and handwritten data.

[0105] Output of judgment results and report generation

[0106] The server generates a report based on the results of the assessment. This report includes the conclusion that "this article was created by generative AI," as well as specific writing style patterns and key phrases. Specifically, the report is generated using Python's ReportLab and Jinja2. The report is generated in HTML or PDF format and can be viewed by the user.

[0107] Notification of results

[0108] The server notifies the user of the generated report. Possible notification methods include sending an email (using the SMTP protocol) or displaying it on a dedicated dashboard (a front-end can be configured using React or Vue.js). Users can check the dashboard to see the results of the assessment of the content they posted.

[0109] Specific examples

[0110] Media company use cases

[0111] A reporter (user) submits a new article to the system from their device. The server receives the article data and cleans the text and removes HTML tags. The server then performs text analysis using the NLTK library, inputs the features into the generative AI judgment algorithm, and finally provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[0112] Example of use at an advertising agency

[0113] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data and extracts text from images and videos using the Tesseract library. The server then analyzes the extracted text using the NLTK library and inputs the features into a judgment algorithm. The final judgment result is that "this ad was not created by generative AI." The ad creator confirms the result via email or dashboard.

[0114] Prompt Sentence Examples

[0115] Prompt: "Determine if the following ad content was created by a generative AI: (text of ad content)"

[0116] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0117] Step 1:

[0118] Receiving content

[0119] A user sends content from a device to a server. For example, a journalist at a media company posts a new article. The input is the content submitted by the user, such as an article, advertisement, or review, and the output is the raw content data stored on the server. The server receives the HTTP request, authenticates the user, and stores the content in a database.

[0120] Step 2:

[0121] Data Preprocessing

[0122] The server preprocesses the received content and extracts the text data. Specifically, it takes the received content data as input, cleans it by removing HTML tags and line breaks, and outputs pure text data. For example, it uses a library such as BeautifulSoup to remove HTML tags and extract the text. If the content contains images or videos, it uses TesseractOCR to extract the characters.

[0123] Step 3:

[0124] Feature extraction

[0125] The server extracts features from the preprocessed text data. The input is the text data obtained in step 2, and the output is a set of features. Specifically, it counts word frequency using the NLTK library and analyzes writing style and grammatical structure using SpaCy. It also vectorizes the text using Word2Vec and extracts features.

[0126] Step 4:

[0127] Execution of generative AI decision algorithm

[0128] The server inputs the extracted features into the generative AI judgment algorithm. The input is the set of features obtained in step 3, and the output is the judgment result of whether the content was created by generative AI. Specifically, the features are input into a pre-trained multilayer neural network model (using TensorFlow or PyTorch) and a judgment is made. This determines whether the content was created by generative AI.

[0129] Step 5:

[0130] Output of judgment results and report generation

[0131] The server generates a report based on the judgment results. The input is the judgment results obtained in step 4, and the output is a report summarizing the judgment results. Specifically, using Python's ReportLab or Jinja2, a report containing the judgment results and their rationale is created in HTML or PDF format. This report includes information such as characteristic writing style patterns and frequently occurring words.

[0132] Step 6:

[0133] Notification of results

[0134] The server notifies the user of the generated report. The input is the report generated in step 5, and the output is a notification email sent to the user or information displayed on a dashboard. Specifically, the server sends an email using the SMTP protocol or displays the information on a front-end dashboard using React or Vue.js. The user receives this notification and checks the results through the dashboard.

[0135] (Application example 1)

[0136] 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."

[0137] While recent advances in generative AI technology have made it easier to generate content, it has also become more difficult to distinguish between content created by humans and content created by generative AI. This can lead to distrust among consumers, particularly in areas such as advertising. Furthermore, when advertisements are automatically generated by generative AI, verification takes time and effort, which can lead to problems with the reliability of the advertisements. Therefore, a system is needed that allows users who view advertisements to verify their authenticity in real time.

[0138] 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.

[0139] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for inputting the extracted features into a generative AI determination algorithm to obtain a determination result, means for acquiring an image of the advertisement using a camera of a specific device (smart glasses) and extracting text in real time, and means for visually displaying the determination result on the display of the device in real time. This enables a user viewing an advertisement to determine in real time whether the content of the advertisement is created by a generative AI and immediately confirm its reliability.

[0140] "Means for receiving content from users" refers to the function of sending user-generated digital content such as articles, advertisements, and reviews to the system and receiving it.

[0141] The "means for preprocessing received content and extracting text data" is a function for removing unnecessary data and noise from received content and converting it into analyzable text data.

[0142] "Means for extracting features from preprocessed text data" is a function for extracting features such as specific patterns and frequencies from cleaned text data.

[0143] "Means of inputting extracted features into a generative AI judgment algorithm to obtain a judgment result" refers to a function that passes the features to a judgment algorithm and uses that algorithm to identify whether the text was created by a generative AI.

[0144] "Means for notifying the user of the judgment result" refers to a function for notifying the user of the judgment result, and includes email, display on the dashboard, etc.

[0145] "Means for obtaining images of advertisements using the camera of a specific device (smart glasses)" refers to a function for capturing visual data of advertisements using the camera installed in the smart glasses.

[0146] "Means for extracting text in real time" is a function that quickly extracts text data from captured images.

[0147] "Means for visually displaying the judgment results on the device display in real time" refers to a function that displays the judgment results on the smart glasses display on the spot, allowing the user to check the results immediately.

[0148] The present invention describes a system for determining whether advertising content received from a user was created by a generation AI in real time.

[0149] First, a user wears smart glasses and views an advertisement. The smart glasses are equipped with a camera that captures an image of the advertisement. The image data captured by the camera is then processed by a computer inside the smart glasses.

[0150] This computer has software (e.g., pytesseract) installed that utilizes optical character recognition (OCR) technology. The OCR software extracts text data from the captured advertisement images. The text data obtained through this OCR process is sent to a server.

[0151] Next, the server preprocesses the received text data to remove unnecessary data and noise. From the preprocessed text data, the server extracts features. These features include word frequency and contextual patterns in the text. Once feature extraction is complete, they are input into a generative AI judgment algorithm. The generative AI judgment algorithm uses a pre-trained model (e.g., a multilayer neural network).

[0152] The server then obtains a determination result as to whether the ad content was generated by AI. This determination result is fed back to the smart glasses in real time. Specifically, the determination result is visually displayed on the smart glasses' display. For example, a message such as "This ad was generated by AI" is displayed.

[0153] This system allows users to check in real time whether the ad they are viewing was created by a human or by a generation AI. This allows them to instantly determine the reliability of the ad and reduces distrust. It also helps the advertising industry effectively prevent the delivery of fraudulent ads by generation AI.

[0154] As a concrete example, the prompt sentence is shown below.

[0155] "Determine whether this ad copy: 'New smartphone, now half price!' was created by generative AI."

[0156] Using this prompt, the generative AI model can analyze the content of the ad copy and provide an appropriate judgment result.

[0157] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0158] Step 1:

[0159] The camera on the smart glasses captures the image of the advertisement that the user is viewing. The input is the advertisement image captured by the smart glasses camera, and the output is the captured advertisement image data. This camera has high resolution and can clearly capture the text in the advertisement.

[0160] Step 2:

[0161] The captured image data is processed by the computer in the smart glasses, and text data is extracted using OCR software (e.g., pytesseract). The input is the advertising image data, and the output is the extracted text data. Here, the OCR software recognizes the characters in the image and generates the analysis results as text.

[0162] Step 3:

[0163] The extracted text data is sent from the smart glasses to a server, where the input is the text data and the output is the text data sent to the server, where the data is transferred to the server through a network interface.

[0164] Step 4:

[0165] The server preprocesses the received text data to remove unnecessary data and noise. The input is raw data and the output is cleaned text data. This preprocessing includes spell checking and filtering of irrelevant content.

[0166] Step 5:

[0167] Features are extracted from the preprocessed text data. The input is cleaned text data, and the output is feature data. The server analyzes this data and quantifies the frequency of specific patterns and words.

[0168] Step 6:

[0169] The extracted features are input into the generative AI judgment algorithm. The input is the feature data, and the output is the generative AI's judgment result. The server makes this judgment using a pre-trained multilayer neural network model.

[0170] Step 7:

[0171] The server sends the judgment result to the smart glasses. The input is the judgment result, and the output is the judgment result sent to the smart glasses. The server sends this data to the smart glasses in real time via the network.

[0172] Step 8:

[0173] The smart glasses visually display the received judgment results on the display. The input is the judgment result data, and the output is visual information displayed on the display. The user can check on the display whether the advertising content was generated by the AI.

[0174] 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.

[0175] The present invention relates to a system that accurately judges content created by a generative AI and recognizes and appropriately responds to user emotions. Specific embodiments are described below with specific examples.

[0176] System configuration

[0177] The system includes the following major components:

[0178] 1. Content Reception Method

[0179] 2. Pretreatment Methods

[0180] 3. Feature Extraction Method

[0181] 4. Generative AI Judgment Algorithm

[0182] 5. Judgment result notification means

[0183] 6. Metadata Extraction Methods

[0184] 7. Report Generation Methods

[0185] 8. Emotion Engine

[0186] Program processing

[0187] Receiving content

[0188] Users upload content such as articles, advertisements, reviews, etc. from their devices. For example, a journalist posts a new article, and an advertising creator uploads advertising content.

[0189] Data Preprocessing

[0190] The server receives the content submitted by the user and pre-processes it, extracting text data. If the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, Optical Character Recognition (OCR) techniques are used to extract text from them.

[0191] Feature extraction

[0192] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[0193] Execution of generative AI decision algorithm

[0194] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[0195] Output of judgment results

[0196] Based on the results, the server determines whether the content was generated by AI. This result also includes information on which features were important.

[0197] Generate reports

[0198] The server generates a report based on its findings, which includes information on whether the content was created by generative AI, the justification for this, and metadata, if necessary. For example, the report could conclude, "This article was not created by generative AI."

[0199] Notification of results

[0200] The server notifies the user of the generated report by email, displaying it on a dashboard, etc. The user can access the dashboard to check the results.

[0201] Use of emotion engine

[0202] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, and text tone to recognize emotions. For example, it can detect if the user is feeling stressed.

[0203] Responding according to emotions

[0204] The server adjusts the notification method based on the emotions recognized by the emotion engine. For example, if the user is feeling stressed, the server can make notifications softer or reduce the frequency of notifications. The content of the report can also be adjusted according to the user's emotions. For example, if the user is feeling anxious, the server can provide additional information to alleviate that anxiety.

[0205] Specific examples

[0206] Media company use cases

[0207] A reporter (user) submits a new article to the system. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides the judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[0208] Example of use at an advertising agency

[0209] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or via dashboard. At the same time, the emotion engine analyzes the ad creator's emotions, and the content of notifications and reports is optimized according to their emotions.

[0210] The above is a specific embodiment of a generative AI content determination system that combines an emotion engine. This not only enables generative AI to identify content, but also provides highly reliable information that takes into account the user's emotions.

[0211] The processing flow will be explained below.

[0212] Step 1:

[0213] Users upload content from their devices, for example, journalists post new articles and advertising creators upload advertising content.

[0214] Step 2:

[0215] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[0216] Step 3:

[0217] The content received by the server is preprocessed. During this preprocessing, text data is extracted. If the content is in a specific format such as HTML or PDF, only the text portion is extracted and cleaned up. Unnecessary tags and special characters are removed.

[0218] Step 4:

[0219] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[0220] Step 5:

[0221] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[0222] Step 6:

[0223] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and stylistic patterns. Specific stylistic tendencies and unnatural phrasing are also detected here.

[0224] Step 7:

[0225] The server inputs the extracted features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI, such as a neural network or decision tree model.

[0226] Step 8:

[0227] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was generated by generative AI, including information on the features used.

[0228] Step 9:

[0229] The server generates a report based on the results of the judgment. This report includes the AI's judgment, its rationale, and metadata, if necessary. For example, the report may conclude, "This article was not created by a generative AI."

[0230] Step 10:

[0231] The server notifies the user of the generated report. Notification methods include sending an email or displaying the report on a dashboard. The user can access the dashboard to check the results.

[0232] Step 11:

[0233] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions. For example, it analyzes whether the user is feeling stressed.

[0234] Step 12:

[0235] The server adjusts the notification method and report content based on the user's emotions recognized by the emotion engine. If an emotion is recognized, the notification can be made softer or the frequency of notifications can be reduced. For example, if the user is feeling anxious, additional information can be provided to alleviate the anxiety.

[0236] These are the specific processing steps for combining an emotion engine with a generative AI content assessment system, which enables the generative AI to identify content with high accuracy and provides highly reliable information that takes into consideration the user's emotions.

[0237] Example 2

[0238] 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."

[0239] Current generative AI assessment systems not only assess content created by generative AI, but also lack the functionality to appropriately notify users of the assessment results and take their feelings into consideration. As a result, they are unable to reduce user stress and anxiety, which could lead to a decline in system satisfaction.

[0240] 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.

[0241] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result, means for generating a report based on the judgment result, means for notifying the user of the report, and means including an emotion engine for recognizing the user's emotions and adjusting the notification method according to the user's emotions. This enables the generative AI to identify content and to provide notifications that take the user's emotions into consideration.

[0242] The "content receiving means" is a function for receiving content such as articles, advertisements, and reviews from users.

[0243] The "preprocessing means" is a function that removes unnecessary information from the received content and extracts text data.

[0244] The "feature extraction means" is a function that extracts features such as word frequency, sentence structure, and style patterns from preprocessed text data.

[0245] The "generative AI determination algorithm" is an algorithm that uses extracted features to determine whether content was created by generative AI.

[0246] The "judgment result notification means" is a function that generates a report based on the obtained judgment result and notifies the user of it.

[0247] An "emotion engine" is a system that analyzes facial expressions, voice, and text tone to recognize a user's emotions.

[0248] The "means for adjusting the notification method according to emotion" is a function for adjusting the notification method based on the emotion of the user detected by the emotion engine.

[0249] The "metadata extraction means" is a function that extracts metadata from content and uses that information for analysis.

[0250] The present invention is a system that accurately judges content created by a generative AI, and further includes a function to recognize the user's emotions and respond appropriately. Specific embodiments of the system are described below.

[0251] Content reception means

[0252] Users upload content such as articles, advertisements, and reviews from their devices. To do so, users access the system's web interface using their PC or smartphone, select the article file, and click the upload button. Specific examples include a journalist posting a new article, or an advertising creator uploading new advertising content.

[0253] Pretreatment means

[0254] The server receives the content sent by the user and performs preprocessing. This process involves removing unnecessary tags from the received content and extracting clean text data. Specifically, it extracts text from HTML content and extracts text from images and videos using optical character recognition (OCR) technology.

[0255] Feature extraction method

[0256] The server extracts features from the preprocessed text data. This function is achieved by using text analysis tools to extract features such as word frequency, sentence structure, and stylistic patterns. For example, it detects the frequent occurrence of unusual words and grammatical unnaturalness, and stores this information in a database.

[0257] Generative AI Judgment Algorithm

[0258] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained multilayer neural network model to determine whether the content was created by generative AI. For example, by inputting the features into the algorithm, it can give a score such as "high possibility of being created by generative AI."

[0259] Judgment result notification means

[0260] The server generates a report based on the judgment result and notifies the user. In this process, a report is created that includes the judgment result, the importance of the features, and detailed analysis information, and notifies the user. For example, a PDF report is generated that concludes, "This article was not created by generative AI," and is sent to the user's email address or displayed on the dashboard.

[0261] Emotion Engine

[0262] The server uses an emotion engine to recognize the user's emotions. This engine has the ability to recognize emotions by analyzing the user's facial expressions, voice, and text tone. For example, when a reporter is reviewing a report, it can detect "anxiety" or "stress" from the user's facial expressions and voice.

[0263] Notification method adjustment method according to emotions

[0264] The server adjusts the notification method based on the user's emotions recognized by the emotion engine. If the user is feeling stressed, the server can soften the notification text and adjust the report content to reassure the user. For example, the server can provide additional information such as "If you need more information about the results, please contact support."

[0265] For example, when a media company posts an article, the server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into a generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[0266] Examples of prompts:

[0267] "How likely is it that this content was created by generative AI?"

[0268] In this way, the present invention is a system that uses generative AI to identify content and adjusts the delivery and notification methods of highly reliable information that takes into account the user's emotions.

[0269] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0270] Step 1: Receiving content

[0271] A user uploads content files such as articles, advertisements, and reviews from a terminal.

[0272] Users access the system's web interface using their PC or smartphone, select the article file, and click the upload button.

[0273] Input: Various content files (text files, HTML files, image files, etc.)

[0274] Output: Content file saved on the server

[0275] Step 2: Preprocessing the data

[0276] The server receives the content sent by the user and performs pre-processing.

[0277] The server removes unnecessary tags from the received HTML file and extracts clean text data.

[0278] For image files, the server uses optical character recognition (OCR) technology to extract the text.

[0279] Input: Saved content file

[0280] Output: Cleaned text data

[0281] Step 3: Feature extraction

[0282] The server extracts features from the preprocessed text data.

[0283] The server uses text analysis tools to identify features such as word frequency, sentence structure, and stylistic patterns.

[0284] Input: Preprocessed text data

[0285] Output: Extracted features (word frequency, sentence structure, stylistic patterns, etc.)

[0286] Step 4: Run the generative AI decision algorithm

[0287] The server inputs the extracted features into the generative AI judgment algorithm.

[0288] The server uses a pre-trained multi-layer neural network model to determine whether the content was created by generative AI.

[0289] Input: extracted features

[0290] Output: Judgment results (score and evaluation) by the generating AI

[0291] Step 5: Output of judgment results

[0292] The server determines whether the content was generated by a generation AI based on the results of the generation AI determination algorithm.

[0293] The server compiles the results of the assessment and the reasons for it in a report.

[0294] Input: Results of the generative AI decision algorithm

[0295] Output: Report containing the judgement results

[0296] Step 6: Generate reports

[0297] The server generates a detailed report based on the results of the assessment.

[0298] The report includes details of the judgment results, the reasons for the judgment, and the results of the feature analysis.

[0299] Input: Judgment result and its reasoning

[0300] Output: Detailed report

[0301] Step 7: Notification of results

[0302] The server notifies the user of the generated report.

[0303] The server sends the report to the user's email address or displays it in a dashboard that the user accesses.

[0304] Input: Generated report

[0305] Output: Email notification or dashboard notification

[0306] Step 8: Use the Emotion Engine

[0307] The server uses an emotion engine to recognize the user's emotions.

[0308] The server analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions.

[0309] Input: User facial expressions, tone of voice, and text

[0310] Output: User emotion recognition results (e.g., anxiety, stress, joy, etc.)

[0311] Step 9: Respond emotionally

[0312] The server adjusts the notification method based on the user's emotion recognized by the emotion engine.

[0313] The server changes the notification wording to softer expressions depending on the emotion and provides additional information as needed.

[0314] Input: User emotion recognition results

[0315] Output: Adjusted notification wording and additional information

[0316] (Application example 2)

[0317] 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."

[0318] Recently, the amount of content created by generative AI has been increasing, and there is a demand for systems that can accurately distinguish between them. However, it is important not only to simply distinguish between AI-generated content, but also to provide optimal information by taking into consideration the emotions of the user viewing the content. However, current systems have insufficient emotion recognition capabilities, which means they are unable to improve the quality of the user experience.

[0319] 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 receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generation AI judgment algorithm to obtain a judgment result, means for recognizing the user's emotions using the smartphone's camera and microphone, means for optimizing the content to be displayed based on the user's emotions, and means for notifying the user of the judgment result. This not only enables the generation AI to accurately distinguish content, but also enables the provision of appropriate content according to the user's emotions.

[0320] "Generative AI" is a system that uses artificial intelligence technology to automatically generate creative content in the same way that humans do.

[0321] "Content" is a collective term for digital information such as text, images, video, and audio that is created and shared by users.

[0322] "Means for receiving" refers to a mechanism, device, or program for receiving content sent from a user at a server or terminal.

[0323] The "means for preprocessing and extracting text data" is a function for unifying the data format of received content and performing processing to extract the necessary text data.

[0324] "Means for extracting features" refers to a method or device for identifying and quantifying useful patterns and attributes from preprocessed text data.

[0325] A "generative AI determination algorithm" is an algorithm that determines whether the input features were generated by a generative AI based on specific rules or models.

[0326] "Means for recognizing emotions" refers to devices or software that analyze the user's voice and facial expressions and estimate their emotional state.

[0327] The "means for optimizing the content to be displayed" is a function for adjusting the content and order of the content to be displayed according to the emotional state of the user.

[0328] System configuration

[0329] This invention is a system that recognizes user emotions and accurately judges content created by generative AI. The system includes the following main components:

[0330] 1. Content Reception Method

[0331] Users upload content such as articles, videos, and audio from their devices. For example, this is how users post news articles.

[0332] 2. Pretreatment Methods

[0333] The server receives the content submitted by the user and pre-processes it to extract text data. For example, if the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, optical character recognition (OCR) techniques are used to extract text from them.

[0334] 3. Feature Extraction Method

[0335] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[0336] 4. Generative AI Judgment Algorithm

[0337] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[0338] 5. Emotion recognition means

[0339] The server uses the smartphone's camera and microphone to recognize the user's emotions. This includes technology that analyzes the user's facial expressions and voice to recognize emotions. It uses facial recognition APIs such as Google's MediaPipe and Azure's Emotion API.

[0340] 6. Feed optimization methods

[0341] The server optimizes the content feed displayed to the user based on the data obtained from the emotion recognition means. For example, if the user is feeling stressed, it will prioritize displaying relaxing content.

[0342] 7. Judgment result notification means

[0343] The server notifies the user of the results. Notification methods include sending an email or displaying the results on the dashboard. The user can access the dashboard to check the results.

[0344] Examples and prompts

[0345] Below is a concrete example of how the system actually works.

[0346] Example: Checking news articles

[0347] Suppose a user reads a new news article on their smartphone. First, the smartphone's microphone records the audio data, and the facial expression camera captures the user's facial expressions. These data are sent to the server, where the emotion recognition means recognizes the emotions. An example of a prompt sentence in this case is as follows:

[0348] Example prompt sentence:

[0349] Determine if the following article was created with generative AI:

[0350] "We bring you the latest economic news about..." (article content)

[0351] The server analyzes the user's emotional state based on the emotion recognition results. Then, a preprocessing means extracts the article's text data and generates features. The generated features are input into a generative AI judgment algorithm to determine whether the article was created by generative AI. Finally, the judgment result is notified to the user, and the feed is optimized based on the user's emotions.

[0352] In this way, users can verify the authenticity of an article and at the same time be provided with appropriate information that matches their emotions at the time.

[0353] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0354] Step 1:

[0355] A user inputs content using a terminal. For example, a user posts a news article on a smartphone. The input content is sent to a server via an application on the smartphone. This is the content receiving means.

[0356] Step 2:

[0357] The server receives content sent by the user using the content receiving means. It then analyzes the content using the preprocessing means and extracts text data. For example, it extracts text from HTML content and performs cleanup processing. It also extracts text from images and videos using OCR technology as needed. The input is the content, and the output is organized text data.

[0358] Step 3:

[0359] The server extracts features from the preprocessed text data. Features include word frequency, sentence structure, and stylistic patterns. This allows specific patterns and unnatural expressions in the text to be quantified. The input is the preprocessed text data, and the output is feature data.

[0360] Step 4:

[0361] The server uses a generative AI judgment algorithm to determine whether the content was created by generative AI based on the extracted features. For example, a pre-trained machine learning model such as a multilayer neural network or decision tree is used. The input is the feature data, and the output is the judgment result.

[0362] Step 5:

[0363] The server recognizes the user's emotions using the smartphone's camera and microphone. The camera captures the user's facial expressions and analyzes them using a facial expression recognition API (such as Google's MediaPipe). The server also collects audio data using the microphone and performs audio emotion analysis. The input is the user's facial expression data and audio data, and the output is the emotion recognition results.

[0364] Step 6:

[0365] The server optimizes the content feed to be displayed to the user based on the emotion recognition results. For example, if the user is feeling stressed, it will prioritize displaying relaxing content. The input is the emotion recognition results, and the output is the optimized content feed.

[0366] Step 7:

[0367] The server notifies the user of the results of the generated AI's judgment. Notification methods include displaying the results on the dashboard or sending an email. The user can access the dashboard and check the contents. The input is the judgment result and feed optimization result, and the output is a user notification.

[0368] 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.

[0369] 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.

[0370] 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.

[0371] [Second embodiment]

[0372] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0373] 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.

[0374] 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).

[0375] 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.

[0376] 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.

[0377] 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).

[0378] 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.

[0379] 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.

[0380] 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.

[0381] 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.

[0382] 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.

[0383] 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."

[0384] The present invention relates to a system that accurately judges content created by generative AI, and specific embodiments thereof will be described below with specific examples.

[0385] System configuration

[0386] This system focuses on receiving content sent by users, preprocessing it to extract text data, inputting features from the text data into a generative AI judgment algorithm, and determining whether the content was created by generative AI.

[0387] Program processing

[0388] Receiving content

[0389] When a user sends content such as an article, advertisement, or review from a terminal, the server receives the content. For example, when a reporter from a media company posts a new article to the system, the article data is sent to the server.

[0390] Data Preprocessing

[0391] The server preprocesses the received content and extracts text data. Preprocessing includes text cleaning and parsing. If the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) to extract text from them.

[0392] Feature extraction

[0393] From the preprocessed text data, the server extracts features, which include word frequency, sentence structure, specific patterns of writing style, etc. Specific features are selected based on a machine learning model.

[0394] Execution of generative AI decision algorithm

[0395] The server inputs the extracted features into a generative AI determination algorithm, which uses a pre-trained model to determine whether the content was created by generative AI. For example, the server uses a machine learning model such as a multilayer neural network or a decision tree to make the determination.

[0396] Output of judgment results and report generation

[0397] The server generates a report based on the results of its assessment. This report includes whether the content was created by a generative AI and the reasons for this. For example, if it concludes that "this article was created by a generative AI," it will cite specific stylistic patterns and unnatural phrasing as reasons for this.

[0398] Notification of results

[0399] The server notifies the user of the generated report. Possible notification methods include sending an email or displaying it on a dashboard. For example, a reporter from a media company can check the dashboard to confirm that the article he or she posted was not generated by a generative AI.

[0400] Specific examples

[0401] Media company use cases

[0402] A reporter (user) submits a new article to the system from their device. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[0403] Example of use at an advertising agency

[0404] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or other means and can confirm them.

[0405] The above is a specific embodiment for carrying out the present invention. By using this system, content can be identified by the generation AI with high accuracy, and the reliability of the information can be ensured.

[0406] The processing flow will be explained below.

[0407] Step 1:

[0408] Users upload content from their devices. For example, journalists post articles and advertising creators upload advertising content.

[0409] Step 2:

[0410] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[0411] Step 3:

[0412] The content received by the server is preprocessed. This preprocessing involves text extraction. If the content is in a specific format like HTML or PDF, only the text data is extracted and cleaned up. Unnecessary tags and special characters are removed.

[0413] Step 4:

[0414] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[0415] Step 5:

[0416] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[0417] Step 6:

[0418] The server generates features from the extracted text data, including word frequency, sentence structure, and specific phrasing patterns. Stylistic patterns and unnatural expressions in the text are also analyzed here.

[0419] Step 7:

[0420] The server inputs the generated features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI.

[0421] Step 8:

[0422] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was created by generative AI, including information on which features were important.

[0423] Step 9:

[0424] The server generates a report based on the findings, including the findings, rationale, and optional metadata, such as a conclusion like "This article was not created by a generative AI."

[0425] Step 10:

[0426] The server notifies the user of the generated report, which can be done by email or displayed on a dashboard. The user can then access the dashboard to check the results.

[0427] These are the specific processing steps of the content assessment system using generative AI. At each step, the server, device, and user play their respective roles, and the system as a whole functions to increase the reliability of information.

[0428] Example 1

[0429] 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."

[0430] Conventional systems have low accuracy in determining content created by generative AI, making it difficult to obtain reliable results. Furthermore, when content is provided in a variety of formats (text, images, video), there is a lack of a way to consistently analyze them, making it difficult to make a comprehensive determination. This has made it difficult to ensure the reliability of information.

[0431] 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.

[0432] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for generating a report based on the judgment result, means for notifying the user of the judgment result, and means for using character recognition technology to extract text from images or videos. This enables highly accurate analysis of various types of content and highly reliable content judgment by the generation AI.

[0433] "Content" refers to information such as articles, advertisements, reviews, etc. submitted by users, and may include various formats such as text, images, and video.

[0434] A "user" is a person or organization that uses this system to send content and receives the results of the evaluation.

[0435] A "server" is a computer system that receives content and performs processes such as preprocessing, feature extraction, execution of a determination algorithm, report generation, and notification of results.

[0436] "Preprocessing" refers to the process of extracting text data from received content and performing cleaning and analysis.

[0437] "Text Data" means textual information extracted from Content, including text with HTML tags removed and text extracted from images or video using OCR technology.

[0438] "Features" are data characteristics extracted from preprocessed text data, and include word frequency, sentence structure, and specific patterns of writing style.

[0439] A "generative AI determination algorithm" is an algorithm that uses a pre-trained model to determine whether content was created by generative AI.

[0440] A "report" is a document that describes the judgment results and their rationale, and is generated in HTML or PDF format.

[0441] "Optical Character Recognition (OCR)" is a technology that extracts character data from images and videos.

[0442] "Judgment result" is the output of the content judgment obtained by the generative AI judgment algorithm.

[0443] "Notification" refers to informing the user of the judgment result, and includes methods such as sending an email or displaying it on a dashboard.

[0444] The present invention relates to a system that judges content created by generative AI with high accuracy, and will be explained below with specific examples.

[0445] Receiving content

[0446] When a user sends content such as an article, advertisement, or review from their device, the server receives the content. For example, when a journalist from a media company posts a new article to the system, the article data is sent to the server. The server authenticates the user and verifies the source of the content. Specifically, it parses the JSON-formatted data packet and extracts the article ID and poster ID.

[0447] Data Preprocessing

[0448] The server preprocesses the received content and extracts text data. Preprocessing includes cleaning and parsing the text. For example, if the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) technology to extract text. Specifically, it uses the Python Tesseract library to extract characters from images.

[0449] Feature extraction

[0450] The server extracts features from the preprocessed text data. These include word frequency, sentence structure, and stylistic patterns. The server uses the NLTK library to count frequently occurring words in the text and encode the words (Word2Vec). It also performs POS tagging and dependency structure analysis to detect stylistic patterns. Specifically, it uses the SpaCy library to perform morphological analysis of sentences and extract important grammatical structures as features.

[0451] Execution of generative AI decision algorithm

[0452] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained model to determine whether the content was created by generative AI. For example, it inputs the features and performs classification using a multi-layer neural network (using TensorFlow or PyTorch). This classification model is previously trained on data from generative AI and handwritten data.

[0453] Output of judgment results and report generation

[0454] The server generates a report based on the results of the assessment. This report includes the conclusion that "this article was created by generative AI," as well as specific writing style patterns and key phrases. Specifically, the report is generated using Python's ReportLab and Jinja2. The report is generated in HTML or PDF format and can be viewed by the user.

[0455] Notification of results

[0456] The server notifies the user of the generated report. Possible notification methods include sending an email (using the SMTP protocol) or displaying it on a dedicated dashboard (a front-end can be configured using React or Vue.js). Users can check the dashboard to see the results of the assessment of the content they posted.

[0457] Specific examples

[0458] Media company use cases

[0459] A reporter (user) submits a new article to the system from their device. The server receives the article data and cleans the text and removes HTML tags. The server then performs text analysis using the NLTK library, inputs the features into the generative AI judgment algorithm, and finally provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[0460] Example of use at an advertising agency

[0461] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data and extracts text from images and videos using the Tesseract library. The server then analyzes the extracted text using the NLTK library and inputs the features into a judgment algorithm. The final judgment result is that "this ad was not created by generative AI." The ad creator confirms the result via email or dashboard.

[0462] Prompt Sentence Examples

[0463] Prompt: "Determine if the following ad content was created by a generative AI: (text of ad content)"

[0464] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0465] Step 1:

[0466] Receiving content

[0467] A user sends content from a device to a server. For example, a journalist at a media company posts a new article. The input is the content submitted by the user, such as an article, advertisement, or review, and the output is the raw content data stored on the server. The server receives the HTTP request, authenticates the user, and stores the content in a database.

[0468] Step 2:

[0469] Data Preprocessing

[0470] The server preprocesses the received content and extracts the text data. Specifically, it takes the received content data as input, cleans it by removing HTML tags and line breaks, and outputs pure text data. For example, it uses a library such as BeautifulSoup to remove HTML tags and extract the text. If the content contains images or videos, it uses TesseractOCR to extract the characters.

[0471] Step 3:

[0472] Feature extraction

[0473] The server extracts features from the preprocessed text data. The input is the text data obtained in step 2, and the output is a set of features. Specifically, it counts word frequency using the NLTK library and analyzes writing style and grammatical structure using SpaCy. It also vectorizes the text using Word2Vec and extracts features.

[0474] Step 4:

[0475] Execution of generative AI decision algorithm

[0476] The server inputs the extracted features into the generative AI judgment algorithm. The input is the set of features obtained in step 3, and the output is the judgment result of whether the content was created by generative AI. Specifically, the features are input into a pre-trained multilayer neural network model (using TensorFlow or PyTorch) and a judgment is made. This determines whether the content was created by generative AI.

[0477] Step 5:

[0478] Output of judgment results and report generation

[0479] The server generates a report based on the judgment results. The input is the judgment results obtained in step 4, and the output is a report summarizing the judgment results. Specifically, using Python's ReportLab or Jinja2, a report containing the judgment results and their rationale is created in HTML or PDF format. This report includes information such as characteristic writing style patterns and frequently occurring words.

[0480] Step 6:

[0481] Notification of results

[0482] The server notifies the user of the generated report. The input is the report generated in step 5, and the output is a notification email sent to the user or information displayed on a dashboard. Specifically, the server sends an email using the SMTP protocol or displays the information on a front-end dashboard using React or Vue.js. The user receives this notification and checks the results through the dashboard.

[0483] (Application example 1)

[0484] 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."

[0485] While recent advances in generative AI technology have made it easier to generate content, it has also become more difficult to distinguish between content created by humans and content created by generative AI. This can lead to distrust among consumers, particularly in areas such as advertising. Furthermore, when advertisements are automatically generated by generative AI, verification takes time and effort, which can lead to problems with the reliability of the advertisements. Therefore, a system is needed that allows users who view advertisements to verify their authenticity in real time.

[0486] 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.

[0487] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for inputting the extracted features into a generative AI determination algorithm to obtain a determination result, means for acquiring an image of the advertisement using a camera of a specific device (smart glasses) and extracting text in real time, and means for visually displaying the determination result on the display of the device in real time. This enables a user viewing an advertisement to determine in real time whether the content of the advertisement is created by a generative AI and immediately confirm its reliability.

[0488] "Means for receiving content from users" refers to the function of sending user-generated digital content such as articles, advertisements, and reviews to the system and receiving it.

[0489] The "means for preprocessing received content and extracting text data" is a function for removing unnecessary data and noise from received content and converting it into analyzable text data.

[0490] "Means for extracting features from preprocessed text data" is a function for extracting features such as specific patterns and frequencies from cleaned text data.

[0491] "Means of inputting extracted features into a generative AI judgment algorithm to obtain a judgment result" refers to a function that passes the features to a judgment algorithm and uses that algorithm to identify whether the text was created by a generative AI.

[0492] "Means for notifying the user of the judgment result" refers to a function for notifying the user of the judgment result, and includes email, display on the dashboard, etc.

[0493] "Means for obtaining images of advertisements using the camera of a specific device (smart glasses)" refers to a function for capturing visual data of advertisements using the camera installed in the smart glasses.

[0494] "Means for extracting text in real time" is a function that quickly extracts text data from captured images.

[0495] "Means for visually displaying the judgment results on the device display in real time" refers to a function that displays the judgment results on the smart glasses display on the spot, allowing the user to check the results immediately.

[0496] The present invention describes a system for determining whether advertising content received from a user was created by a generation AI in real time.

[0497] First, a user wears smart glasses and views an advertisement. The smart glasses are equipped with a camera that captures an image of the advertisement. The image data captured by the camera is then processed by a computer inside the smart glasses.

[0498] This computer has software (e.g., pytesseract) installed that utilizes optical character recognition (OCR) technology. The OCR software extracts text data from the captured advertisement images. The text data obtained through this OCR process is sent to a server.

[0499] Next, the server preprocesses the received text data to remove unnecessary data and noise. From the preprocessed text data, the server extracts features. These features include word frequency and contextual patterns in the text. Once feature extraction is complete, they are input into a generative AI judgment algorithm. The generative AI judgment algorithm uses a pre-trained model (e.g., a multilayer neural network).

[0500] The server then obtains a determination result as to whether the ad content was generated by AI. This determination result is fed back to the smart glasses in real time. Specifically, the determination result is visually displayed on the smart glasses' display. For example, a message such as "This ad was generated by AI" is displayed.

[0501] This system allows users to check in real time whether the ad they are viewing was created by a human or by a generation AI. This allows them to instantly determine the reliability of the ad and reduces distrust. It also helps the advertising industry effectively prevent the delivery of fraudulent ads by generation AI.

[0502] As a concrete example, the prompt sentence is shown below.

[0503] "Determine whether this ad copy: 'New smartphone, now half price!' was created by generative AI."

[0504] Using this prompt, the generative AI model can analyze the content of the ad copy and provide an appropriate judgment result.

[0505] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0506] Step 1:

[0507] The camera on the smart glasses captures the image of the advertisement that the user is viewing. The input is the advertisement image captured by the smart glasses camera, and the output is the captured advertisement image data. This camera has high resolution and can clearly capture the text in the advertisement.

[0508] Step 2:

[0509] The captured image data is processed by the computer in the smart glasses, and text data is extracted using OCR software (e.g., pytesseract). The input is the advertising image data, and the output is the extracted text data. Here, the OCR software recognizes the characters in the image and generates the analysis results as text.

[0510] Step 3:

[0511] The extracted text data is sent from the smart glasses to a server, where the input is the text data and the output is the text data sent to the server, where the data is transferred to the server through a network interface.

[0512] Step 4:

[0513] The server preprocesses the received text data to remove unnecessary data and noise. The input is raw data and the output is cleaned text data. This preprocessing includes spell checking and filtering of irrelevant content.

[0514] Step 5:

[0515] Features are extracted from the preprocessed text data. The input is cleaned text data, and the output is feature data. The server analyzes this data and quantifies the frequency of specific patterns and words.

[0516] Step 6:

[0517] The extracted features are input into the generative AI judgment algorithm. The input is the feature data, and the output is the generative AI's judgment result. The server makes this judgment using a pre-trained multilayer neural network model.

[0518] Step 7:

[0519] The server sends the judgment result to the smart glasses. The input is the judgment result, and the output is the judgment result sent to the smart glasses. The server sends this data to the smart glasses in real time via the network.

[0520] Step 8:

[0521] The smart glasses visually display the received judgment results on the display. The input is the judgment result data, and the output is visual information displayed on the display. The user can check on the display whether the advertising content was generated by the AI.

[0522] 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.

[0523] The present invention relates to a system that accurately judges content created by a generative AI and recognizes and appropriately responds to user emotions. Specific embodiments are described below with specific examples.

[0524] System configuration

[0525] The system includes the following major components:

[0526] 1. Content Reception Method

[0527] 2. Pretreatment Methods

[0528] 3. Feature Extraction Method

[0529] 4. Generative AI Judgment Algorithm

[0530] 5. Judgment result notification means

[0531] 6. Metadata Extraction Methods

[0532] 7. Report Generation Methods

[0533] 8. Emotion Engine

[0534] Program processing

[0535] Receiving content

[0536] Users upload content such as articles, advertisements, reviews, etc. from their devices. For example, a journalist posts a new article, and an advertising creator uploads advertising content.

[0537] Data Preprocessing

[0538] The server receives the content submitted by the user and pre-processes it, extracting text data. If the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, Optical Character Recognition (OCR) techniques are used to extract text from them.

[0539] Feature extraction

[0540] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[0541] Execution of generative AI decision algorithm

[0542] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[0543] Output of judgment results

[0544] Based on the results, the server determines whether the content was generated by AI. This result also includes information on which features were important.

[0545] Generate reports

[0546] The server generates a report based on its findings, which includes information on whether the content was created by generative AI, the justification for this, and metadata, if necessary. For example, the report could conclude, "This article was not created by generative AI."

[0547] Notification of results

[0548] The server notifies the user of the generated report by email, displaying it on a dashboard, etc. The user can access the dashboard to check the results.

[0549] Use of emotion engine

[0550] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, and text tone to recognize emotions. For example, it can detect if the user is feeling stressed.

[0551] Responding according to emotions

[0552] The server adjusts the notification method based on the emotions recognized by the emotion engine. For example, if the user is feeling stressed, the server can make notifications softer or reduce the frequency of notifications. The content of the report can also be adjusted according to the user's emotions. For example, if the user is feeling anxious, the server can provide additional information to alleviate that anxiety.

[0553] Specific examples

[0554] Media company use cases

[0555] A reporter (user) submits a new article to the system. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides the judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[0556] Example of use at an advertising agency

[0557] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or via dashboard. At the same time, the emotion engine analyzes the ad creator's emotions, and the content of notifications and reports is optimized according to their emotions.

[0558] The above is a specific embodiment of a generative AI content determination system that combines an emotion engine. This not only enables generative AI to identify content, but also provides highly reliable information that takes into account the user's emotions.

[0559] The processing flow will be explained below.

[0560] Step 1:

[0561] Users upload content from their devices, for example, journalists post new articles and advertising creators upload advertising content.

[0562] Step 2:

[0563] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[0564] Step 3:

[0565] The content received by the server is preprocessed. During this preprocessing, text data is extracted. If the content is in a specific format such as HTML or PDF, only the text portion is extracted and cleaned up. Unnecessary tags and special characters are removed.

[0566] Step 4:

[0567] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[0568] Step 5:

[0569] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[0570] Step 6:

[0571] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and stylistic patterns. Specific stylistic tendencies and unnatural phrasing are also detected here.

[0572] Step 7:

[0573] The server inputs the extracted features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI, such as a neural network or decision tree model.

[0574] Step 8:

[0575] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was generated by generative AI, including information on the features used.

[0576] Step 9:

[0577] The server generates a report based on the results of the judgment. This report includes the AI's judgment, its rationale, and metadata, if necessary. For example, the report may conclude, "This article was not created by a generative AI."

[0578] Step 10:

[0579] The server notifies the user of the generated report. Notification methods include sending an email or displaying the report on a dashboard. The user can access the dashboard to check the results.

[0580] Step 11:

[0581] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions. For example, it analyzes whether the user is feeling stressed.

[0582] Step 12:

[0583] The server adjusts the notification method and report content based on the user's emotions recognized by the emotion engine. If an emotion is recognized, the notification can be made softer or the frequency of notifications can be reduced. For example, if the user is feeling anxious, additional information can be provided to alleviate the anxiety.

[0584] These are the specific processing steps for combining an emotion engine with a generative AI content assessment system, which enables the generative AI to identify content with high accuracy and provides highly reliable information that takes into consideration the user's emotions.

[0585] Example 2

[0586] 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."

[0587] Current generative AI assessment systems not only assess content created by generative AI, but also lack the functionality to appropriately notify users of the assessment results and take their feelings into consideration. As a result, they are unable to reduce user stress and anxiety, which could lead to a decline in system satisfaction.

[0588] 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.

[0589] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result, means for generating a report based on the judgment result, means for notifying the user of the report, and means including an emotion engine for recognizing the user's emotions and adjusting the notification method according to the user's emotions. This enables the generative AI to identify content and to provide notifications that take the user's emotions into consideration.

[0590] The "content receiving means" is a function for receiving content such as articles, advertisements, and reviews from users.

[0591] The "preprocessing means" is a function that removes unnecessary information from the received content and extracts text data.

[0592] The "feature extraction means" is a function that extracts features such as word frequency, sentence structure, and style patterns from preprocessed text data.

[0593] The "generative AI determination algorithm" is an algorithm that uses extracted features to determine whether content was created by generative AI.

[0594] The "judgment result notification means" is a function that generates a report based on the obtained judgment result and notifies the user of it.

[0595] An "emotion engine" is a system that analyzes facial expressions, voice, and text tone to recognize a user's emotions.

[0596] The "means for adjusting the notification method according to emotion" is a function for adjusting the notification method based on the emotion of the user detected by the emotion engine.

[0597] The "metadata extraction means" is a function that extracts metadata from content and uses that information for analysis.

[0598] The present invention is a system that accurately judges content created by a generative AI, and further includes a function to recognize the user's emotions and respond appropriately. Specific embodiments of the system are described below.

[0599] Content reception means

[0600] Users upload content such as articles, advertisements, and reviews from their devices. To do so, users access the system's web interface using their PC or smartphone, select the article file, and click the upload button. Specific examples include a journalist posting a new article, or an advertising creator uploading new advertising content.

[0601] Pretreatment means

[0602] The server receives the content sent by the user and performs preprocessing. This process involves removing unnecessary tags from the received content and extracting clean text data. Specifically, it extracts text from HTML content and extracts text from images and videos using optical character recognition (OCR) technology.

[0603] Feature extraction method

[0604] The server extracts features from the preprocessed text data. This function is achieved by using text analysis tools to extract features such as word frequency, sentence structure, and stylistic patterns. For example, it detects the frequent occurrence of unusual words and grammatical unnaturalness, and stores this information in a database.

[0605] Generative AI Judgment Algorithm

[0606] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained multilayer neural network model to determine whether the content was created by generative AI. For example, by inputting the features into the algorithm, it can give a score such as "high possibility of being created by generative AI."

[0607] Judgment result notification means

[0608] The server generates a report based on the judgment result and notifies the user. In this process, a report is created that includes the judgment result, the importance of the features, and detailed analysis information, and notifies the user. For example, a PDF report is generated that concludes, "This article was not created by generative AI," and is sent to the user's email address or displayed on the dashboard.

[0609] Emotion Engine

[0610] The server uses an emotion engine to recognize the user's emotions. This engine has the ability to recognize emotions by analyzing the user's facial expressions, voice, and text tone. For example, when a reporter is reviewing a report, it can detect "anxiety" or "stress" from the user's facial expressions and voice.

[0611] Notification method adjustment method according to emotions

[0612] The server adjusts the notification method based on the user's emotions recognized by the emotion engine. If the user is feeling stressed, the server can soften the notification text and adjust the report content to reassure the user. For example, the server can provide additional information such as "If you need more information about the results, please contact support."

[0613] For example, when a media company posts an article, the server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into a generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[0614] Examples of prompts:

[0615] "How likely is it that this content was created by generative AI?"

[0616] In this way, the present invention is a system that uses generative AI to identify content and adjusts the delivery and notification methods of highly reliable information that takes into account the user's emotions.

[0617] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0618] Step 1: Receiving content

[0619] A user uploads content files such as articles, advertisements, and reviews from a terminal.

[0620] Users access the system's web interface using their PC or smartphone, select the article file, and click the upload button.

[0621] Input: Various content files (text files, HTML files, image files, etc.)

[0622] Output: Content file saved on the server

[0623] Step 2: Preprocessing the data

[0624] The server receives the content sent by the user and performs pre-processing.

[0625] The server removes unnecessary tags from the received HTML file and extracts clean text data.

[0626] For image files, the server uses optical character recognition (OCR) technology to extract the text.

[0627] Input: Saved content file

[0628] Output: Cleaned text data

[0629] Step 3: Feature extraction

[0630] The server extracts features from the preprocessed text data.

[0631] The server uses text analysis tools to identify features such as word frequency, sentence structure, and stylistic patterns.

[0632] Input: Preprocessed text data

[0633] Output: Extracted features (word frequency, sentence structure, stylistic patterns, etc.)

[0634] Step 4: Run the generative AI decision algorithm

[0635] The server inputs the extracted features into the generative AI judgment algorithm.

[0636] The server uses a pre-trained multi-layer neural network model to determine whether the content was created by generative AI.

[0637] Input: extracted features

[0638] Output: Judgment results (score and evaluation) by the generating AI

[0639] Step 5: Output of judgment results

[0640] The server determines whether the content was generated by a generation AI based on the results of the generation AI determination algorithm.

[0641] The server compiles the results of the assessment and the reasons for it in a report.

[0642] Input: Results of the generative AI decision algorithm

[0643] Output: Report containing the judgement results

[0644] Step 6: Generate reports

[0645] The server generates a detailed report based on the results of the assessment.

[0646] The report includes details of the judgment results, the reasons for the judgment, and the results of the feature analysis.

[0647] Input: Judgment result and its reasoning

[0648] Output: Detailed report

[0649] Step 7: Notification of results

[0650] The server notifies the user of the generated report.

[0651] The server sends the report to the user's email address or displays it in a dashboard that the user accesses.

[0652] Input: Generated report

[0653] Output: Email notification or dashboard notification

[0654] Step 8: Use the Emotion Engine

[0655] The server uses an emotion engine to recognize the user's emotions.

[0656] The server analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions.

[0657] Input: User facial expressions, tone of voice, and text

[0658] Output: User emotion recognition results (e.g., anxiety, stress, joy, etc.)

[0659] Step 9: Respond emotionally

[0660] The server adjusts the notification method based on the user's emotion recognized by the emotion engine.

[0661] The server changes the notification wording to softer expressions depending on the emotion and provides additional information as needed.

[0662] Input: User emotion recognition results

[0663] Output: Adjusted notification wording and additional information

[0664] (Application example 2)

[0665] 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."

[0666] Recently, the amount of content created by generative AI has been increasing, and there is a demand for systems that can accurately distinguish between them. However, it is important not only to simply distinguish between AI-generated content, but also to provide optimal information by taking into consideration the emotions of the user viewing the content. However, current systems have insufficient emotion recognition capabilities, which means they are unable to improve the quality of the user experience.

[0667] 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 receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generation AI judgment algorithm to obtain a judgment result, means for recognizing the user's emotions using the smartphone's camera and microphone, means for optimizing the content to be displayed based on the user's emotions, and means for notifying the user of the judgment result. This not only enables the generation AI to accurately distinguish content, but also enables the provision of appropriate content according to the user's emotions.

[0668] "Generative AI" is a system that uses artificial intelligence technology to automatically generate creative content in the same way that humans do.

[0669] "Content" is a collective term for digital information such as text, images, video, and audio that is created and shared by users.

[0670] "Means for receiving" refers to a mechanism, device, or program for receiving content sent from a user at a server or terminal.

[0671] The "means for preprocessing and extracting text data" is a function for unifying the data format of received content and performing processing to extract the necessary text data.

[0672] "Means for extracting features" refers to a method or device for identifying and quantifying useful patterns and attributes from preprocessed text data.

[0673] A "generative AI determination algorithm" is an algorithm that determines whether the input features were generated by a generative AI based on specific rules or models.

[0674] "Means for recognizing emotions" refers to devices or software that analyze the user's voice and facial expressions and estimate their emotional state.

[0675] The "means for optimizing the content to be displayed" is a function for adjusting the content and order of the content to be displayed according to the emotional state of the user.

[0676] System configuration

[0677] This invention is a system that recognizes user emotions and accurately judges content created by generative AI. The system includes the following main components:

[0678] 1. Content Reception Method

[0679] Users upload content such as articles, videos, and audio from their devices. For example, this is how users post news articles.

[0680] 2. Pretreatment Methods

[0681] The server receives the content submitted by the user and pre-processes it to extract text data. For example, if the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, optical character recognition (OCR) techniques are used to extract text from them.

[0682] 3. Feature Extraction Method

[0683] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[0684] 4. Generative AI Judgment Algorithm

[0685] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[0686] 5. Emotion recognition means

[0687] The server uses the smartphone's camera and microphone to recognize the user's emotions. This includes technology that analyzes the user's facial expressions and voice to recognize emotions. It uses facial recognition APIs such as Google's MediaPipe and Azure's Emotion API.

[0688] 6. Feed optimization methods

[0689] The server optimizes the content feed displayed to the user based on the data obtained from the emotion recognition means. For example, if the user is feeling stressed, it will prioritize displaying relaxing content.

[0690] 7. Judgment result notification means

[0691] The server notifies the user of the results. Notification methods include sending an email or displaying the results on the dashboard. The user can access the dashboard to check the results.

[0692] Examples and prompts

[0693] Below is a concrete example of how the system actually works.

[0694] Example: Checking news articles

[0695] Suppose a user reads a new news article on their smartphone. First, the smartphone's microphone records the audio data, and the facial expression camera captures the user's facial expressions. These data are sent to the server, where the emotion recognition means recognizes the emotions. An example of a prompt sentence in this case is as follows:

[0696] Example prompt sentence:

[0697] Determine if the following article was created with generative AI:

[0698] "We bring you the latest economic news about..." (article content)

[0699] The server analyzes the user's emotional state based on the emotion recognition results. Then, a preprocessing means extracts the article's text data and generates features. The generated features are input into a generative AI judgment algorithm to determine whether the article was created by generative AI. Finally, the judgment result is notified to the user, and the feed is optimized based on the user's emotions.

[0700] In this way, users can verify the authenticity of an article and at the same time be provided with appropriate information that matches their emotions at the time.

[0701] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0702] Step 1:

[0703] A user inputs content using a terminal. For example, a user posts a news article on a smartphone. The input content is sent to a server via an application on the smartphone. This is the content receiving means.

[0704] Step 2:

[0705] The server receives content sent by the user using the content receiving means. It then analyzes the content using the preprocessing means and extracts text data. For example, it extracts text from HTML content and performs cleanup processing. It also extracts text from images and videos using OCR technology as needed. The input is the content, and the output is organized text data.

[0706] Step 3:

[0707] The server extracts features from the preprocessed text data. Features include word frequency, sentence structure, and stylistic patterns. This allows specific patterns and unnatural expressions in the text to be quantified. The input is the preprocessed text data, and the output is feature data.

[0708] Step 4:

[0709] The server uses a generative AI judgment algorithm to determine whether the content was created by generative AI based on the extracted features. For example, a pre-trained machine learning model such as a multilayer neural network or decision tree is used. The input is the feature data, and the output is the judgment result.

[0710] Step 5:

[0711] The server recognizes the user's emotions using the smartphone's camera and microphone. The camera captures the user's facial expressions and analyzes them using a facial expression recognition API (such as Google's MediaPipe). The server also collects audio data using the microphone and performs audio emotion analysis. The input is the user's facial expression data and audio data, and the output is the emotion recognition results.

[0712] Step 6:

[0713] The server optimizes the content feed to be displayed to the user based on the emotion recognition results. For example, if the user is feeling stressed, it will prioritize displaying relaxing content. The input is the emotion recognition results, and the output is the optimized content feed.

[0714] Step 7:

[0715] The server notifies the user of the results of the generated AI's judgment. Notification methods include displaying the results on the dashboard or sending an email. The user can access the dashboard and check the contents. The input is the judgment result and feed optimization result, and the output is a user notification.

[0716] 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.

[0717] 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.

[0718] 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.

[0719] [Third embodiment]

[0720] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0721] 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.

[0722] 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).

[0723] 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.

[0724] 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.

[0725] 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).

[0726] 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.

[0727] 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.

[0728] 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.

[0729] 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.

[0730] 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.

[0731] 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."

[0732] The present invention relates to a system that accurately judges content created by generative AI, and specific embodiments thereof will be described below with specific examples.

[0733] System configuration

[0734] This system focuses on receiving content sent by users, preprocessing it to extract text data, inputting features from the text data into a generative AI judgment algorithm, and determining whether the content was created by generative AI.

[0735] Program processing

[0736] Receiving content

[0737] When a user sends content such as an article, advertisement, or review from a terminal, the server receives the content. For example, when a reporter from a media company posts a new article to the system, the article data is sent to the server.

[0738] Data Preprocessing

[0739] The server preprocesses the received content and extracts text data. Preprocessing includes text cleaning and parsing. If the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) to extract text from them.

[0740] Feature extraction

[0741] From the preprocessed text data, the server extracts features, which include word frequency, sentence structure, specific patterns of writing style, etc. Specific features are selected based on a machine learning model.

[0742] Execution of generative AI decision algorithm

[0743] The server inputs the extracted features into a generative AI determination algorithm, which uses a pre-trained model to determine whether the content was created by generative AI. For example, the server uses a machine learning model such as a multilayer neural network or a decision tree to make the determination.

[0744] Output of judgment results and report generation

[0745] The server generates a report based on the results of its assessment. This report includes whether the content was created by a generative AI and the reasons for this. For example, if it concludes that "this article was created by a generative AI," it will cite specific stylistic patterns and unnatural phrasing as reasons for this.

[0746] Notification of results

[0747] The server notifies the user of the generated report. Possible notification methods include sending an email or displaying it on a dashboard. For example, a reporter from a media company can check the dashboard to confirm that the article he or she posted was not generated by a generative AI.

[0748] Specific examples

[0749] Media company use cases

[0750] A reporter (user) submits a new article to the system from their device. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[0751] Example of use at an advertising agency

[0752] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or other means and can confirm them.

[0753] The above is a specific embodiment for carrying out the present invention. By using this system, content can be identified by the generation AI with high accuracy, and the reliability of the information can be ensured.

[0754] The processing flow will be explained below.

[0755] Step 1:

[0756] Users upload content from their devices. For example, journalists post articles and advertising creators upload advertising content.

[0757] Step 2:

[0758] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[0759] Step 3:

[0760] The content received by the server is preprocessed. This preprocessing involves text extraction. If the content is in a specific format like HTML or PDF, only the text data is extracted and cleaned up. Unnecessary tags and special characters are removed.

[0761] Step 4:

[0762] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[0763] Step 5:

[0764] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[0765] Step 6:

[0766] The server generates features from the extracted text data, including word frequency, sentence structure, and specific phrasing patterns. Stylistic patterns and unnatural expressions in the text are also analyzed here.

[0767] Step 7:

[0768] The server inputs the generated features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI.

[0769] Step 8:

[0770] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was created by generative AI, including information on which features were important.

[0771] Step 9:

[0772] The server generates a report based on the findings, including the findings, rationale, and optional metadata, such as a conclusion like "This article was not created by a generative AI."

[0773] Step 10:

[0774] The server notifies the user of the generated report, which can be done by email or displayed on a dashboard. The user can then access the dashboard to check the results.

[0775] These are the specific processing steps of the content assessment system using generative AI. At each step, the server, device, and user play their respective roles, and the system as a whole functions to increase the reliability of information.

[0776] Example 1

[0777] 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."

[0778] Conventional systems have low accuracy in determining content created by generative AI, making it difficult to obtain reliable results. Furthermore, when content is provided in a variety of formats (text, images, video), there is a lack of a way to consistently analyze them, making it difficult to make a comprehensive determination. This has made it difficult to ensure the reliability of information.

[0779] 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.

[0780] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for generating a report based on the judgment result, means for notifying the user of the judgment result, and means for using character recognition technology to extract text from images or videos. This enables highly accurate analysis of various types of content and highly reliable content judgment by the generation AI.

[0781] "Content" refers to information such as articles, advertisements, reviews, etc. submitted by users, and may include various formats such as text, images, and video.

[0782] A "user" is a person or organization that uses this system to send content and receives the results of the evaluation.

[0783] A "server" is a computer system that receives content and performs processes such as preprocessing, feature extraction, execution of a determination algorithm, report generation, and notification of results.

[0784] "Preprocessing" refers to the process of extracting text data from received content and performing cleaning and analysis.

[0785] "Text Data" means textual information extracted from Content, including text with HTML tags removed and text extracted from images or video using OCR technology.

[0786] "Features" are data characteristics extracted from preprocessed text data, and include word frequency, sentence structure, and specific patterns of writing style.

[0787] A "generative AI determination algorithm" is an algorithm that uses a pre-trained model to determine whether content was created by generative AI.

[0788] A "report" is a document that describes the judgment results and their rationale, and is generated in HTML or PDF format.

[0789] "Optical Character Recognition (OCR)" is a technology that extracts character data from images and videos.

[0790] "Judgment result" is the output of the content judgment obtained by the generative AI judgment algorithm.

[0791] "Notification" refers to informing the user of the judgment result, and includes methods such as sending an email or displaying it on a dashboard.

[0792] The present invention relates to a system that judges content created by generative AI with high accuracy, and will be explained below with specific examples.

[0793] Receiving content

[0794] When a user sends content such as an article, advertisement, or review from their device, the server receives the content. For example, when a journalist from a media company posts a new article to the system, the article data is sent to the server. The server authenticates the user and verifies the source of the content. Specifically, it parses the JSON-formatted data packet and extracts the article ID and poster ID.

[0795] Data Preprocessing

[0796] The server preprocesses the received content and extracts text data. Preprocessing includes cleaning and parsing the text. For example, if the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) technology to extract text. Specifically, it uses the Python Tesseract library to extract characters from images.

[0797] Feature extraction

[0798] The server extracts features from the preprocessed text data. These include word frequency, sentence structure, and stylistic patterns. The server uses the NLTK library to count frequently occurring words in the text and encode the words (Word2Vec). It also performs POS tagging and dependency structure analysis to detect stylistic patterns. Specifically, it uses the SpaCy library to perform morphological analysis of sentences and extract important grammatical structures as features.

[0799] Execution of generative AI decision algorithm

[0800] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained model to determine whether the content was created by generative AI. For example, it inputs the features and performs classification using a multi-layer neural network (using TensorFlow or PyTorch). This classification model is previously trained on data from generative AI and handwritten data.

[0801] Output of judgment results and report generation

[0802] The server generates a report based on the results of the assessment. This report includes the conclusion that "this article was created by generative AI," as well as specific writing style patterns and key phrases. Specifically, the report is generated using Python's ReportLab and Jinja2. The report is generated in HTML or PDF format and can be viewed by the user.

[0803] Notification of results

[0804] The server notifies the user of the generated report. Possible notification methods include sending an email (using the SMTP protocol) or displaying it on a dedicated dashboard (a front-end can be configured using React or Vue.js). Users can check the dashboard to see the results of the assessment of the content they posted.

[0805] Specific examples

[0806] Media company use cases

[0807] A reporter (user) submits a new article to the system from their device. The server receives the article data and cleans the text and removes HTML tags. The server then performs text analysis using the NLTK library, inputs the features into the generative AI judgment algorithm, and finally provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[0808] Example of use at an advertising agency

[0809] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data and extracts text from images and videos using the Tesseract library. The server then analyzes the extracted text using the NLTK library and inputs the features into a judgment algorithm. The final judgment result is that "this ad was not created by generative AI." The ad creator confirms the result via email or dashboard.

[0810] Prompt Sentence Examples

[0811] Prompt: "Determine if the following ad content was created by a generative AI: (text of ad content)"

[0812] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0813] Step 1:

[0814] Receiving content

[0815] A user sends content from a device to a server. For example, a journalist at a media company posts a new article. The input is the content submitted by the user, such as an article, advertisement, or review, and the output is the raw content data stored on the server. The server receives the HTTP request, authenticates the user, and stores the content in a database.

[0816] Step 2:

[0817] Data Preprocessing

[0818] The server preprocesses the received content and extracts the text data. Specifically, it takes the received content data as input, cleans it by removing HTML tags and line breaks, and outputs pure text data. For example, it uses a library such as BeautifulSoup to remove HTML tags and extract the text. If the content contains images or videos, it uses TesseractOCR to extract the characters.

[0819] Step 3:

[0820] Feature extraction

[0821] The server extracts features from the preprocessed text data. The input is the text data obtained in step 2, and the output is a set of features. Specifically, it counts word frequency using the NLTK library and analyzes writing style and grammatical structure using SpaCy. It also vectorizes the text using Word2Vec and extracts features.

[0822] Step 4:

[0823] Execution of generative AI decision algorithm

[0824] The server inputs the extracted features into the generative AI judgment algorithm. The input is the set of features obtained in step 3, and the output is the judgment result of whether the content was created by generative AI. Specifically, the features are input into a pre-trained multilayer neural network model (using TensorFlow or PyTorch) and a judgment is made. This determines whether the content was created by generative AI.

[0825] Step 5:

[0826] Output of judgment results and report generation

[0827] The server generates a report based on the judgment results. The input is the judgment results obtained in step 4, and the output is a report summarizing the judgment results. Specifically, using Python's ReportLab or Jinja2, a report containing the judgment results and their rationale is created in HTML or PDF format. This report includes information such as characteristic writing style patterns and frequently occurring words.

[0828] Step 6:

[0829] Notification of results

[0830] The server notifies the user of the generated report. The input is the report generated in step 5, and the output is a notification email sent to the user or information displayed on a dashboard. Specifically, the server sends an email using the SMTP protocol or displays the information on a front-end dashboard using React or Vue.js. The user receives this notification and checks the results through the dashboard.

[0831] (Application example 1)

[0832] 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."

[0833] While recent advances in generative AI technology have made it easier to generate content, it has also become more difficult to distinguish between content created by humans and content created by generative AI. This can lead to distrust among consumers, particularly in areas such as advertising. Furthermore, when advertisements are automatically generated by generative AI, verification takes time and effort, which can lead to problems with the reliability of the advertisements. Therefore, a system is needed that allows users who view advertisements to verify their authenticity in real time.

[0834] 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.

[0835] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for inputting the extracted features into a generative AI determination algorithm to obtain a determination result, means for acquiring an image of the advertisement using a camera of a specific device (smart glasses) and extracting text in real time, and means for visually displaying the determination result on the display of the device in real time. This enables a user viewing an advertisement to determine in real time whether the content of the advertisement is created by a generative AI and immediately confirm its reliability.

[0836] "Means for receiving content from users" refers to the function of sending user-generated digital content such as articles, advertisements, and reviews to the system and receiving it.

[0837] The "means for preprocessing received content and extracting text data" is a function for removing unnecessary data and noise from received content and converting it into analyzable text data.

[0838] "Means for extracting features from preprocessed text data" is a function for extracting features such as specific patterns and frequencies from cleaned text data.

[0839] "Means of inputting extracted features into a generative AI judgment algorithm to obtain a judgment result" refers to a function that passes the features to a judgment algorithm and uses that algorithm to identify whether the text was created by a generative AI.

[0840] "Means for notifying the user of the judgment result" refers to a function for notifying the user of the judgment result, and includes email, display on the dashboard, etc.

[0841] "Means for obtaining images of advertisements using the camera of a specific device (smart glasses)" refers to a function for capturing visual data of advertisements using the camera installed in the smart glasses.

[0842] "Means for extracting text in real time" is a function that quickly extracts text data from captured images.

[0843] "Means for visually displaying the judgment results on the device display in real time" refers to a function that displays the judgment results on the smart glasses display on the spot, allowing the user to check the results immediately.

[0844] The present invention describes a system for determining whether advertising content received from a user was created by a generation AI in real time.

[0845] First, a user wears smart glasses and views an advertisement. The smart glasses are equipped with a camera that captures an image of the advertisement. The image data captured by the camera is then processed by a computer inside the smart glasses.

[0846] This computer has software (e.g., pytesseract) installed that utilizes optical character recognition (OCR) technology. The OCR software extracts text data from the captured advertisement images. The text data obtained through this OCR process is sent to a server.

[0847] Next, the server preprocesses the received text data to remove unnecessary data and noise. From the preprocessed text data, the server extracts features. These features include word frequency and contextual patterns in the text. Once feature extraction is complete, they are input into a generative AI judgment algorithm. The generative AI judgment algorithm uses a pre-trained model (e.g., a multilayer neural network).

[0848] The server then obtains a determination result as to whether the ad content was generated by AI. This determination result is fed back to the smart glasses in real time. Specifically, the determination result is visually displayed on the smart glasses' display. For example, a message such as "This ad was generated by AI" is displayed.

[0849] This system allows users to check in real time whether the ad they are viewing was created by a human or by a generation AI. This allows them to instantly determine the reliability of the ad and reduces distrust. It also helps the advertising industry effectively prevent the delivery of fraudulent ads by generation AI.

[0850] As a concrete example, the prompt sentence is shown below.

[0851] "Determine whether this ad copy: 'New smartphone, now half price!' was created by generative AI."

[0852] Using this prompt, the generative AI model can analyze the content of the ad copy and provide an appropriate judgment result.

[0853] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0854] Step 1:

[0855] The camera on the smart glasses captures the image of the advertisement that the user is viewing. The input is the advertisement image captured by the smart glasses camera, and the output is the captured advertisement image data. This camera has high resolution and can clearly capture the text in the advertisement.

[0856] Step 2:

[0857] The captured image data is processed by the computer in the smart glasses, and text data is extracted using OCR software (e.g., pytesseract). The input is the advertising image data, and the output is the extracted text data. Here, the OCR software recognizes the characters in the image and generates the analysis results as text.

[0858] Step 3:

[0859] The extracted text data is sent from the smart glasses to a server, where the input is the text data and the output is the text data sent to the server, where the data is transferred to the server through a network interface.

[0860] Step 4:

[0861] The server preprocesses the received text data to remove unnecessary data and noise. The input is raw data and the output is cleaned text data. This preprocessing includes spell checking and filtering of irrelevant content.

[0862] Step 5:

[0863] Features are extracted from the preprocessed text data. The input is cleaned text data, and the output is feature data. The server analyzes this data and quantifies the frequency of specific patterns and words.

[0864] Step 6:

[0865] The extracted features are input into the generative AI judgment algorithm. The input is the feature data, and the output is the generative AI's judgment result. The server makes this judgment using a pre-trained multilayer neural network model.

[0866] Step 7:

[0867] The server sends the judgment result to the smart glasses. The input is the judgment result, and the output is the judgment result sent to the smart glasses. The server sends this data to the smart glasses in real time via the network.

[0868] Step 8:

[0869] The smart glasses visually display the received judgment results on the display. The input is the judgment result data, and the output is visual information displayed on the display. The user can check on the display whether the advertising content was generated by the AI.

[0870] 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.

[0871] The present invention relates to a system that accurately judges content created by a generative AI and recognizes and appropriately responds to user emotions. Specific embodiments are described below with specific examples.

[0872] System configuration

[0873] The system includes the following major components:

[0874] 1. Content Reception Method

[0875] 2. Pretreatment Methods

[0876] 3. Feature Extraction Method

[0877] 4. Generative AI Judgment Algorithm

[0878] 5. Judgment result notification means

[0879] 6. Metadata Extraction Methods

[0880] 7. Report Generation Methods

[0881] 8. Emotion Engine

[0882] Program processing

[0883] Receiving content

[0884] Users upload content such as articles, advertisements, reviews, etc. from their devices. For example, a journalist posts a new article, and an advertising creator uploads advertising content.

[0885] Data Preprocessing

[0886] The server receives the content submitted by the user and pre-processes it, extracting text data. If the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, Optical Character Recognition (OCR) techniques are used to extract text from them.

[0887] Feature extraction

[0888] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[0889] Execution of generative AI decision algorithm

[0890] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[0891] Output of judgment results

[0892] Based on the results, the server determines whether the content was generated by AI. This result also includes information on which features were important.

[0893] Generate reports

[0894] The server generates a report based on its findings, which includes information on whether the content was created by generative AI, the justification for this, and metadata, if necessary. For example, the report could conclude, "This article was not created by generative AI."

[0895] Notification of results

[0896] The server notifies the user of the generated report by email, displaying it on a dashboard, etc. The user can access the dashboard to check the results.

[0897] Use of emotion engine

[0898] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, and text tone to recognize emotions. For example, it can detect if the user is feeling stressed.

[0899] Responding according to emotions

[0900] The server adjusts the notification method based on the emotions recognized by the emotion engine. For example, if the user is feeling stressed, the server can make notifications softer or reduce the frequency of notifications. The content of the report can also be adjusted according to the user's emotions. For example, if the user is feeling anxious, the server can provide additional information to alleviate that anxiety.

[0901] Specific examples

[0902] Media company use cases

[0903] A reporter (user) submits a new article to the system. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides the judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[0904] Example of use at an advertising agency

[0905] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or via dashboard. At the same time, the emotion engine analyzes the ad creator's emotions, and the content of notifications and reports is optimized according to their emotions.

[0906] The above is a specific embodiment of a generative AI content determination system that combines an emotion engine. This not only enables generative AI to identify content, but also provides highly reliable information that takes into account the user's emotions.

[0907] The processing flow will be explained below.

[0908] Step 1:

[0909] Users upload content from their devices, for example, journalists post new articles and advertising creators upload advertising content.

[0910] Step 2:

[0911] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[0912] Step 3:

[0913] The content received by the server is preprocessed. During this preprocessing, text data is extracted. If the content is in a specific format such as HTML or PDF, only the text portion is extracted and cleaned up. Unnecessary tags and special characters are removed.

[0914] Step 4:

[0915] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[0916] Step 5:

[0917] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[0918] Step 6:

[0919] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and stylistic patterns. Specific stylistic tendencies and unnatural phrasing are also detected here.

[0920] Step 7:

[0921] The server inputs the extracted features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI, such as a neural network or decision tree model.

[0922] Step 8:

[0923] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was generated by generative AI, including information on the features used.

[0924] Step 9:

[0925] The server generates a report based on the results of the judgment. This report includes the AI's judgment, its rationale, and metadata, if necessary. For example, the report may conclude, "This article was not created by a generative AI."

[0926] Step 10:

[0927] The server notifies the user of the generated report. Notification methods include sending an email or displaying the report on a dashboard. The user can access the dashboard to check the results.

[0928] Step 11:

[0929] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions. For example, it analyzes whether the user is feeling stressed.

[0930] Step 12:

[0931] The server adjusts the notification method and report content based on the user's emotions recognized by the emotion engine. If an emotion is recognized, the notification can be made softer or the frequency of notifications can be reduced. For example, if the user is feeling anxious, additional information can be provided to alleviate the anxiety.

[0932] These are the specific processing steps for combining an emotion engine with a generative AI content assessment system, which enables the generative AI to identify content with high accuracy and provides highly reliable information that takes into consideration the user's emotions.

[0933] Example 2

[0934] 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."

[0935] Current generative AI assessment systems not only assess content created by generative AI, but also lack the functionality to appropriately notify users of the assessment results and take their feelings into consideration. As a result, they are unable to reduce user stress and anxiety, which could lead to a decline in system satisfaction.

[0936] 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.

[0937] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result, means for generating a report based on the judgment result, means for notifying the user of the report, and means including an emotion engine for recognizing the user's emotions and adjusting the notification method according to the user's emotions. This enables the generative AI to identify content and to provide notifications that take the user's emotions into consideration.

[0938] The "content receiving means" is a function for receiving content such as articles, advertisements, and reviews from users.

[0939] The "preprocessing means" is a function that removes unnecessary information from the received content and extracts text data.

[0940] The "feature extraction means" is a function that extracts features such as word frequency, sentence structure, and style patterns from preprocessed text data.

[0941] The "generative AI determination algorithm" is an algorithm that uses extracted features to determine whether content was created by generative AI.

[0942] The "judgment result notification means" is a function that generates a report based on the obtained judgment result and notifies the user of it.

[0943] An "emotion engine" is a system that analyzes facial expressions, voice, and text tone to recognize a user's emotions.

[0944] The "means for adjusting the notification method according to emotion" is a function for adjusting the notification method based on the emotion of the user detected by the emotion engine.

[0945] The "metadata extraction means" is a function that extracts metadata from content and uses that information for analysis.

[0946] The present invention is a system that accurately judges content created by a generative AI, and further includes a function to recognize the user's emotions and respond appropriately. Specific embodiments of the system are described below.

[0947] Content reception means

[0948] Users upload content such as articles, advertisements, and reviews from their devices. To do so, users access the system's web interface using their PC or smartphone, select the article file, and click the upload button. Specific examples include a journalist posting a new article, or an advertising creator uploading new advertising content.

[0949] Pretreatment means

[0950] The server receives the content sent by the user and performs preprocessing. This process involves removing unnecessary tags from the received content and extracting clean text data. Specifically, it extracts text from HTML content and extracts text from images and videos using optical character recognition (OCR) technology.

[0951] Feature extraction method

[0952] The server extracts features from the preprocessed text data. This function is achieved by using text analysis tools to extract features such as word frequency, sentence structure, and stylistic patterns. For example, it detects the frequent occurrence of unusual words and grammatical unnaturalness, and stores this information in a database.

[0953] Generative AI Judgment Algorithm

[0954] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained multilayer neural network model to determine whether the content was created by generative AI. For example, by inputting the features into the algorithm, it can give a score such as "high possibility of being created by generative AI."

[0955] Judgment result notification means

[0956] The server generates a report based on the judgment result and notifies the user. In this process, a report is created that includes the judgment result, the importance of the features, and detailed analysis information, and notifies the user. For example, a PDF report is generated that concludes, "This article was not created by generative AI," and is sent to the user's email address or displayed on the dashboard.

[0957] Emotion Engine

[0958] The server uses an emotion engine to recognize the user's emotions. This engine has the ability to recognize emotions by analyzing the user's facial expressions, voice, and text tone. For example, when a reporter is reviewing a report, it can detect "anxiety" or "stress" from the user's facial expressions and voice.

[0959] Notification method adjustment method according to emotions

[0960] The server adjusts the notification method based on the user's emotions recognized by the emotion engine. If the user is feeling stressed, the server can soften the notification text and adjust the report content to reassure the user. For example, the server can provide additional information such as "If you need more information about the results, please contact support."

[0961] For example, when a media company posts an article, the server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into a generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[0962] Examples of prompts:

[0963] "How likely is it that this content was created by generative AI?"

[0964] In this way, the present invention is a system that uses generative AI to identify content and adjusts the delivery and notification methods of highly reliable information that takes into account the user's emotions.

[0965] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0966] Step 1: Receiving content

[0967] A user uploads content files such as articles, advertisements, and reviews from a terminal.

[0968] Users access the system's web interface using their PC or smartphone, select the article file, and click the upload button.

[0969] Input: Various content files (text files, HTML files, image files, etc.)

[0970] Output: Content file saved on the server

[0971] Step 2: Preprocessing the data

[0972] The server receives the content sent by the user and performs pre-processing.

[0973] The server removes unnecessary tags from the received HTML file and extracts clean text data.

[0974] For image files, the server uses optical character recognition (OCR) technology to extract the text.

[0975] Input: Saved content file

[0976] Output: Cleaned text data

[0977] Step 3: Feature extraction

[0978] The server extracts features from the preprocessed text data.

[0979] The server uses text analysis tools to identify features such as word frequency, sentence structure, and stylistic patterns.

[0980] Input: Preprocessed text data

[0981] Output: Extracted features (word frequency, sentence structure, stylistic patterns, etc.)

[0982] Step 4: Run the generative AI decision algorithm

[0983] The server inputs the extracted features into the generative AI judgment algorithm.

[0984] The server uses a pre-trained multi-layer neural network model to determine whether the content was created by generative AI.

[0985] Input: extracted features

[0986] Output: Judgment results (score and evaluation) by the generating AI

[0987] Step 5: Output of judgment results

[0988] The server determines whether the content was generated by a generation AI based on the results of the generation AI determination algorithm.

[0989] The server compiles the results of the assessment and the reasons for it in a report.

[0990] Input: Results of the generative AI decision algorithm

[0991] Output: Report containing the judgement results

[0992] Step 6: Generate reports

[0993] The server generates a detailed report based on the results of the assessment.

[0994] The report includes details of the judgment results, the reasons for the judgment, and the results of the feature analysis.

[0995] Input: Judgment result and its reasoning

[0996] Output: Detailed report

[0997] Step 7: Notification of results

[0998] The server notifies the user of the generated report.

[0999] The server sends the report to the user's email address or displays it in a dashboard that the user accesses.

[1000] Input: Generated report

[1001] Output: Email notification or dashboard notification

[1002] Step 8: Use the Emotion Engine

[1003] The server uses an emotion engine to recognize the user's emotions.

[1004] The server analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions.

[1005] Input: User facial expressions, tone of voice, and text

[1006] Output: User emotion recognition results (e.g., anxiety, stress, joy, etc.)

[1007] Step 9: Respond emotionally

[1008] The server adjusts the notification method based on the user's emotion recognized by the emotion engine.

[1009] The server changes the notification wording to softer expressions depending on the emotion and provides additional information as needed.

[1010] Input: User emotion recognition results

[1011] Output: Adjusted notification wording and additional information

[1012] (Application example 2)

[1013] 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."

[1014] Recently, the amount of content created by generative AI has been increasing, and there is a demand for systems that can accurately distinguish between them. However, it is important not only to simply distinguish between AI-generated content, but also to provide optimal information by taking into consideration the emotions of the user viewing the content. However, current systems have insufficient emotion recognition capabilities, which means they are unable to improve the quality of the user experience.

[1015] 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 receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generation AI judgment algorithm to obtain a judgment result, means for recognizing the user's emotions using the smartphone's camera and microphone, means for optimizing the content to be displayed based on the user's emotions, and means for notifying the user of the judgment result. This not only enables the generation AI to accurately distinguish content, but also enables the provision of appropriate content according to the user's emotions.

[1016] "Generative AI" is a system that uses artificial intelligence technology to automatically generate creative content in the same way that humans do.

[1017] "Content" is a collective term for digital information such as text, images, video, and audio that is created and shared by users.

[1018] "Means for receiving" refers to a mechanism, device, or program for receiving content sent from a user at a server or terminal.

[1019] The "means for preprocessing and extracting text data" is a function for unifying the data format of received content and performing processing to extract the necessary text data.

[1020] "Means for extracting features" refers to a method or device for identifying and quantifying useful patterns and attributes from preprocessed text data.

[1021] A "generative AI determination algorithm" is an algorithm that determines whether the input features were generated by a generative AI based on specific rules or models.

[1022] "Means for recognizing emotions" refers to devices or software that analyze the user's voice and facial expressions and estimate their emotional state.

[1023] The "means for optimizing the content to be displayed" is a function for adjusting the content and order of the content to be displayed according to the emotional state of the user.

[1024] System configuration

[1025] This invention is a system that recognizes user emotions and accurately judges content created by generative AI. The system includes the following main components:

[1026] 1. Content Reception Method

[1027] Users upload content such as articles, videos, and audio from their devices. For example, this is how users post news articles.

[1028] 2. Pretreatment Methods

[1029] The server receives the content submitted by the user and pre-processes it to extract text data. For example, if the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, optical character recognition (OCR) techniques are used to extract text from them.

[1030] 3. Feature Extraction Method

[1031] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[1032] 4. Generative AI Judgment Algorithm

[1033] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[1034] 5. Emotion recognition means

[1035] The server uses the smartphone's camera and microphone to recognize the user's emotions. This includes technology that analyzes the user's facial expressions and voice to recognize emotions. It uses facial recognition APIs such as Google's MediaPipe and Azure's Emotion API.

[1036] 6. Feed optimization methods

[1037] The server optimizes the content feed displayed to the user based on the data obtained from the emotion recognition means. For example, if the user is feeling stressed, it will prioritize displaying relaxing content.

[1038] 7. Judgment result notification means

[1039] The server notifies the user of the results. Notification methods include sending an email or displaying the results on the dashboard. The user can access the dashboard to check the results.

[1040] Examples and prompts

[1041] Below is a concrete example of how the system actually works.

[1042] Example: Checking news articles

[1043] Suppose a user reads a new news article on their smartphone. First, the smartphone's microphone records the audio data, and the facial expression camera captures the user's facial expressions. These data are sent to the server, where the emotion recognition means recognizes the emotions. An example of a prompt sentence in this case is as follows:

[1044] Example prompt sentence:

[1045] Determine if the following article was created with generative AI:

[1046] "We bring you the latest economic news about..." (article content)

[1047] The server analyzes the user's emotional state based on the emotion recognition results. Then, a preprocessing means extracts the article's text data and generates features. The generated features are input into a generative AI judgment algorithm to determine whether the article was created by generative AI. Finally, the judgment result is notified to the user, and the feed is optimized based on the user's emotions.

[1048] In this way, users can verify the authenticity of an article and at the same time be provided with appropriate information that matches their emotions at the time.

[1049] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1050] Step 1:

[1051] A user inputs content using a terminal. For example, a user posts a news article on a smartphone. The input content is sent to a server via an application on the smartphone. This is the content receiving means.

[1052] Step 2:

[1053] The server receives content sent by the user using the content receiving means. It then analyzes the content using the preprocessing means and extracts text data. For example, it extracts text from HTML content and performs cleanup processing. It also extracts text from images and videos using OCR technology as needed. The input is the content, and the output is organized text data.

[1054] Step 3:

[1055] The server extracts features from the preprocessed text data. Features include word frequency, sentence structure, and stylistic patterns. This allows specific patterns and unnatural expressions in the text to be quantified. The input is the preprocessed text data, and the output is feature data.

[1056] Step 4:

[1057] The server uses a generative AI judgment algorithm to determine whether the content was created by generative AI based on the extracted features. For example, a pre-trained machine learning model such as a multilayer neural network or decision tree is used. The input is the feature data, and the output is the judgment result.

[1058] Step 5:

[1059] The server recognizes the user's emotions using the smartphone's camera and microphone. The camera captures the user's facial expressions and analyzes them using a facial expression recognition API (such as Google's MediaPipe). The server also collects audio data using the microphone and performs audio emotion analysis. The input is the user's facial expression data and audio data, and the output is the emotion recognition results.

[1060] Step 6:

[1061] The server optimizes the content feed to be displayed to the user based on the emotion recognition results. For example, if the user is feeling stressed, it will prioritize displaying relaxing content. The input is the emotion recognition results, and the output is the optimized content feed.

[1062] Step 7:

[1063] The server notifies the user of the results of the generated AI's judgment. Notification methods include displaying the results on the dashboard or sending an email. The user can access the dashboard and check the contents. The input is the judgment result and feed optimization result, and the output is a user notification.

[1064] 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.

[1065] 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.

[1066] 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.

[1067] [Fourth embodiment]

[1068] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1069] 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.

[1070] 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).

[1071] 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.

[1072] 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.

[1073] 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).

[1074] 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.

[1075] 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.

[1076] 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.

[1077] 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.

[1078] 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.

[1079] 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.

[1080] 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."

[1081] The present invention relates to a system that accurately judges content created by generative AI, and specific embodiments thereof will be described below with specific examples.

[1082] System configuration

[1083] This system focuses on receiving content sent by users, preprocessing it to extract text data, inputting features from the text data into a generative AI judgment algorithm, and determining whether the content was created by generative AI.

[1084] Program processing

[1085] Receiving content

[1086] When a user sends content such as an article, advertisement, or review from a terminal, the server receives the content. For example, when a reporter from a media company posts a new article to the system, the article data is sent to the server.

[1087] Data Preprocessing

[1088] The server preprocesses the received content and extracts text data. Preprocessing includes text cleaning and parsing. If the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) to extract text from them.

[1089] Feature extraction

[1090] From the preprocessed text data, the server extracts features, which include word frequency, sentence structure, specific patterns of writing style, etc. Specific features are selected based on a machine learning model.

[1091] Execution of generative AI decision algorithm

[1092] The server inputs the extracted features into a generative AI determination algorithm, which uses a pre-trained model to determine whether the content was created by generative AI. For example, the server uses a machine learning model such as a multilayer neural network or a decision tree to make the determination.

[1093] Output of judgment results and report generation

[1094] The server generates a report based on the results of its assessment. This report includes whether the content was created by a generative AI and the reasons for this. For example, if it concludes that "this article was created by a generative AI," it will cite specific stylistic patterns and unnatural phrasing as reasons for this.

[1095] Notification of results

[1096] The server notifies the user of the generated report. Possible notification methods include sending an email or displaying it on a dashboard. For example, a reporter from a media company can check the dashboard to confirm that the article he or she posted was not generated by a generative AI.

[1097] Specific examples

[1098] Media company use cases

[1099] A reporter (user) submits a new article to the system from their device. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[1100] Example of use at an advertising agency

[1101] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or other means and can confirm them.

[1102] The above is a specific embodiment for carrying out the present invention. By using this system, content can be identified by the generation AI with high accuracy, and the reliability of the information can be ensured.

[1103] The processing flow will be explained below.

[1104] Step 1:

[1105] Users upload content from their devices. For example, journalists post articles and advertising creators upload advertising content.

[1106] Step 2:

[1107] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[1108] Step 3:

[1109] The content received by the server is preprocessed. This preprocessing involves text extraction. If the content is in a specific format like HTML or PDF, only the text data is extracted and cleaned up. Unnecessary tags and special characters are removed.

[1110] Step 4:

[1111] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[1112] Step 5:

[1113] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[1114] Step 6:

[1115] The server generates features from the extracted text data, including word frequency, sentence structure, and specific phrasing patterns. Stylistic patterns and unnatural expressions in the text are also analyzed here.

[1116] Step 7:

[1117] The server inputs the generated features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI.

[1118] Step 8:

[1119] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was created by generative AI, including information on which features were important.

[1120] Step 9:

[1121] The server generates a report based on the findings, including the findings, rationale, and optional metadata, such as a conclusion like "This article was not created by a generative AI."

[1122] Step 10:

[1123] The server notifies the user of the generated report, which can be done by email or displayed on a dashboard. The user can then access the dashboard to check the results.

[1124] These are the specific processing steps of the content assessment system using generative AI. At each step, the server, device, and user play their respective roles, and the system as a whole functions to increase the reliability of information.

[1125] Example 1

[1126] 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."

[1127] Conventional systems have low accuracy in determining content created by generative AI, making it difficult to obtain reliable results. Furthermore, when content is provided in a variety of formats (text, images, video), there is a lack of a way to consistently analyze them, making it difficult to make a comprehensive determination. This has made it difficult to ensure the reliability of information.

[1128] 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.

[1129] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for generating a report based on the judgment result, means for notifying the user of the judgment result, and means for using character recognition technology to extract text from images or videos. This enables highly accurate analysis of various types of content and highly reliable content judgment by the generation AI.

[1130] "Content" refers to information such as articles, advertisements, reviews, etc. submitted by users, and may include various formats such as text, images, and video.

[1131] A "user" is a person or organization that uses this system to send content and receives the results of the evaluation.

[1132] A "server" is a computer system that receives content and performs processes such as preprocessing, feature extraction, execution of a determination algorithm, report generation, and notification of results.

[1133] "Preprocessing" refers to the process of extracting text data from received content and performing cleaning and analysis.

[1134] "Text Data" means textual information extracted from Content, including text with HTML tags removed and text extracted from images or video using OCR technology.

[1135] "Features" are data characteristics extracted from preprocessed text data, and include word frequency, sentence structure, and specific patterns of writing style.

[1136] A "generative AI determination algorithm" is an algorithm that uses a pre-trained model to determine whether content was created by generative AI.

[1137] A "report" is a document that describes the judgment results and their rationale, and is generated in HTML or PDF format.

[1138] "Optical Character Recognition (OCR)" is a technology that extracts character data from images and videos.

[1139] "Judgment result" is the output of the content judgment obtained by the generative AI judgment algorithm.

[1140] "Notification" refers to informing the user of the judgment result, and includes methods such as sending an email or displaying it on a dashboard.

[1141] The present invention relates to a system that judges content created by generative AI with high accuracy, and will be explained below with specific examples.

[1142] Receiving content

[1143] When a user sends content such as an article, advertisement, or review from their device, the server receives the content. For example, when a journalist from a media company posts a new article to the system, the article data is sent to the server. The server authenticates the user and verifies the source of the content. Specifically, it parses the JSON-formatted data packet and extracts the article ID and poster ID.

[1144] Data Preprocessing

[1145] The server preprocesses the received content and extracts text data. Preprocessing includes cleaning and parsing the text. For example, if the content is in HTML format, the server extracts the text and removes unnecessary tags. If the content contains images or videos, the server uses optical character recognition (OCR) technology to extract text. Specifically, it uses the Python Tesseract library to extract characters from images.

[1146] Feature extraction

[1147] The server extracts features from the preprocessed text data. These include word frequency, sentence structure, and stylistic patterns. The server uses the NLTK library to count frequently occurring words in the text and encode the words (Word2Vec). It also performs POS tagging and dependency structure analysis to detect stylistic patterns. Specifically, it uses the SpaCy library to perform morphological analysis of sentences and extract important grammatical structures as features.

[1148] Execution of generative AI decision algorithm

[1149] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained model to determine whether the content was created by generative AI. For example, it inputs the features and performs classification using a multi-layer neural network (using TensorFlow or PyTorch). This classification model is previously trained on data from generative AI and handwritten data.

[1150] Output of judgment results and report generation

[1151] The server generates a report based on the results of the assessment. This report includes the conclusion that "this article was created by generative AI," as well as specific writing style patterns and key phrases. Specifically, the report is generated using Python's ReportLab and Jinja2. The report is generated in HTML or PDF format and can be viewed by the user.

[1152] Notification of results

[1153] The server notifies the user of the generated report. Possible notification methods include sending an email (using the SMTP protocol) or displaying it on a dedicated dashboard (a front-end can be configured using React or Vue.js). Users can check the dashboard to see the results of the assessment of the content they posted.

[1154] Specific examples

[1155] Media company use cases

[1156] A reporter (user) submits a new article to the system from their device. The server receives the article data and cleans the text and removes HTML tags. The server then performs text analysis using the NLTK library, inputs the features into the generative AI judgment algorithm, and finally provides a judgment result that "this article was not created by generative AI." The reporter can access the dashboard and check this result.

[1157] Example of use at an advertising agency

[1158] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data and extracts text from images and videos using the Tesseract library. The server then analyzes the extracted text using the NLTK library and inputs the features into a judgment algorithm. The final judgment result is that "this ad was not created by generative AI." The ad creator confirms the result via email or dashboard.

[1159] Prompt Sentence Examples

[1160] Prompt: "Determine if the following ad content was created by a generative AI: (text of ad content)"

[1161] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1162] Step 1:

[1163] Receiving content

[1164] A user sends content from a device to a server. For example, a journalist at a media company posts a new article. The input is the content submitted by the user, such as an article, advertisement, or review, and the output is the raw content data stored on the server. The server receives the HTTP request, authenticates the user, and stores the content in a database.

[1165] Step 2:

[1166] Data Preprocessing

[1167] The server preprocesses the received content and extracts the text data. Specifically, it takes the received content data as input, cleans it by removing HTML tags and line breaks, and outputs pure text data. For example, it uses a library such as BeautifulSoup to remove HTML tags and extract the text. If the content contains images or videos, it uses TesseractOCR to extract the characters.

[1168] Step 3:

[1169] Feature extraction

[1170] The server extracts features from the preprocessed text data. The input is the text data obtained in step 2, and the output is a set of features. Specifically, it counts word frequency using the NLTK library and analyzes writing style and grammatical structure using SpaCy. It also vectorizes the text using Word2Vec and extracts features.

[1171] Step 4:

[1172] Execution of generative AI decision algorithm

[1173] The server inputs the extracted features into the generative AI judgment algorithm. The input is the set of features obtained in step 3, and the output is the judgment result of whether the content was created by generative AI. Specifically, the features are input into a pre-trained multilayer neural network model (using TensorFlow or PyTorch) and a judgment is made. This determines whether the content was created by generative AI.

[1174] Step 5:

[1175] Output of judgment results and report generation

[1176] The server generates a report based on the judgment results. The input is the judgment results obtained in step 4, and the output is a report summarizing the judgment results. Specifically, using Python's ReportLab or Jinja2, a report containing the judgment results and their rationale is created in HTML or PDF format. This report includes information such as characteristic writing style patterns and frequently occurring words.

[1177] Step 6:

[1178] Notification of results

[1179] The server notifies the user of the generated report. The input is the report generated in step 5, and the output is a notification email sent to the user or information displayed on a dashboard. Specifically, the server sends an email using the SMTP protocol or displays the information on a front-end dashboard using React or Vue.js. The user receives this notification and checks the results through the dashboard.

[1180] (Application example 1)

[1181] 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."

[1182] While recent advances in generative AI technology have made it easier to generate content, it has also become more difficult to distinguish between content created by humans and content created by generative AI. This can lead to distrust among consumers, particularly in areas such as advertising. Furthermore, when advertisements are automatically generated by generative AI, verification takes time and effort, which can lead to problems with the reliability of the advertisements. Therefore, a system is needed that allows users who view advertisements to verify their authenticity in real time.

[1183] 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.

[1184] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for inputting the extracted features into a generative AI determination algorithm to obtain a determination result, means for acquiring an image of the advertisement using a camera of a specific device (smart glasses) and extracting text in real time, and means for visually displaying the determination result on the display of the device in real time. This enables a user viewing an advertisement to determine in real time whether the content of the advertisement is created by a generative AI and immediately confirm its reliability.

[1185] "Means for receiving content from users" refers to the function of sending user-generated digital content such as articles, advertisements, and reviews to the system and receiving it.

[1186] The "means for preprocessing received content and extracting text data" is a function for removing unnecessary data and noise from received content and converting it into analyzable text data.

[1187] "Means for extracting features from preprocessed text data" is a function for extracting features such as specific patterns and frequencies from cleaned text data.

[1188] "Means of inputting extracted features into a generative AI judgment algorithm to obtain a judgment result" refers to a function that passes the features to a judgment algorithm and uses that algorithm to identify whether the text was created by a generative AI.

[1189] "Means for notifying the user of the judgment result" refers to a function for notifying the user of the judgment result, and includes email, display on the dashboard, etc.

[1190] "Means for obtaining images of advertisements using the camera of a specific device (smart glasses)" refers to a function for capturing visual data of advertisements using the camera installed in the smart glasses.

[1191] "Means for extracting text in real time" is a function that quickly extracts text data from captured images.

[1192] "Means for visually displaying the judgment results on the device display in real time" refers to a function that displays the judgment results on the smart glasses display on the spot, allowing the user to check the results immediately.

[1193] The present invention describes a system for determining whether advertising content received from a user was created by a generation AI in real time.

[1194] First, a user wears smart glasses and views an advertisement. The smart glasses are equipped with a camera that captures an image of the advertisement. The image data captured by the camera is then processed by a computer inside the smart glasses.

[1195] This computer has software (e.g., pytesseract) installed that utilizes optical character recognition (OCR) technology. The OCR software extracts text data from the captured advertisement images. The text data obtained through this OCR process is sent to a server.

[1196] Next, the server preprocesses the received text data to remove unnecessary data and noise. From the preprocessed text data, the server extracts features. These features include word frequency and contextual patterns in the text. Once feature extraction is complete, they are input into a generative AI judgment algorithm. The generative AI judgment algorithm uses a pre-trained model (e.g., a multilayer neural network).

[1197] The server then obtains a determination result as to whether the ad content was generated by AI. This determination result is fed back to the smart glasses in real time. Specifically, the determination result is visually displayed on the smart glasses' display. For example, a message such as "This ad was generated by AI" is displayed.

[1198] This system allows users to check in real time whether the ad they are viewing was created by a human or by a generation AI. This allows them to instantly determine the reliability of the ad and reduces distrust. It also helps the advertising industry effectively prevent the delivery of fraudulent ads by generation AI.

[1199] As a concrete example, the prompt sentence is shown below.

[1200] "Determine whether this ad copy: 'New smartphone, now half price!' was created by generative AI."

[1201] Using this prompt, the generative AI model can analyze the content of the ad copy and provide an appropriate judgment result.

[1202] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1203] Step 1:

[1204] The camera on the smart glasses captures the image of the advertisement that the user is viewing. The input is the advertisement image captured by the smart glasses camera, and the output is the captured advertisement image data. This camera has high resolution and can clearly capture the text in the advertisement.

[1205] Step 2:

[1206] The captured image data is processed by the computer in the smart glasses, and text data is extracted using OCR software (e.g., pytesseract). The input is the advertising image data, and the output is the extracted text data. Here, the OCR software recognizes the characters in the image and generates the analysis results as text.

[1207] Step 3:

[1208] The extracted text data is sent from the smart glasses to a server, where the input is the text data and the output is the text data sent to the server, where the data is transferred to the server through a network interface.

[1209] Step 4:

[1210] The server preprocesses the received text data to remove unnecessary data and noise. The input is raw data and the output is cleaned text data. This preprocessing includes spell checking and filtering of irrelevant content.

[1211] Step 5:

[1212] Features are extracted from the preprocessed text data. The input is cleaned text data, and the output is feature data. The server analyzes this data and quantifies the frequency of specific patterns and words.

[1213] Step 6:

[1214] The extracted features are input into the generative AI judgment algorithm. The input is the feature data, and the output is the generative AI's judgment result. The server makes this judgment using a pre-trained multilayer neural network model.

[1215] Step 7:

[1216] The server sends the judgment result to the smart glasses. The input is the judgment result, and the output is the judgment result sent to the smart glasses. The server sends this data to the smart glasses in real time via the network.

[1217] Step 8:

[1218] The smart glasses visually display the received judgment results on the display. The input is the judgment result data, and the output is visual information displayed on the display. The user can check on the display whether the advertising content was generated by the AI.

[1219] 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.

[1220] The present invention relates to a system that accurately judges content created by a generative AI and recognizes and appropriately responds to user emotions. Specific embodiments are described below with specific examples.

[1221] System configuration

[1222] The system includes the following major components:

[1223] 1. Content Reception Method

[1224] 2. Pretreatment Methods

[1225] 3. Feature Extraction Method

[1226] 4. Generative AI Judgment Algorithm

[1227] 5. Judgment result notification means

[1228] 6. Metadata Extraction Methods

[1229] 7. Report Generation Methods

[1230] 8. Emotion Engine

[1231] Program processing

[1232] Receiving content

[1233] Users upload content such as articles, advertisements, reviews, etc. from their devices. For example, a journalist posts a new article, and an advertising creator uploads advertising content.

[1234] Data Preprocessing

[1235] The server receives the content submitted by the user and pre-processes it, extracting text data. If the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, Optical Character Recognition (OCR) techniques are used to extract text from them.

[1236] Feature extraction

[1237] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[1238] Execution of generative AI decision algorithm

[1239] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[1240] Output of judgment results

[1241] Based on the results, the server determines whether the content was generated by AI. This result also includes information on which features were important.

[1242] Generate reports

[1243] The server generates a report based on its findings, which includes information on whether the content was created by generative AI, the justification for this, and metadata, if necessary. For example, the report could conclude, "This article was not created by generative AI."

[1244] Notification of results

[1245] The server notifies the user of the generated report by email, displaying it on a dashboard, etc. The user can access the dashboard to check the results.

[1246] Use of emotion engine

[1247] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, and text tone to recognize emotions. For example, it can detect if the user is feeling stressed.

[1248] Responding according to emotions

[1249] The server adjusts the notification method based on the emotions recognized by the emotion engine. For example, if the user is feeling stressed, the server can make notifications softer or reduce the frequency of notifications. The content of the report can also be adjusted according to the user's emotions. For example, if the user is feeling anxious, the server can provide additional information to alleviate that anxiety.

[1250] Specific examples

[1251] Media company use cases

[1252] A reporter (user) submits a new article to the system. The server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into the generative AI judgment algorithm, which ultimately provides the judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[1253] Example of use at an advertising agency

[1254] The ad creator (user) uploads new ad content to the system from their device. The server receives the ad data, extracts text from images and videos, and generates features. These features are then input into the generative AI judgment algorithm, which ultimately determines that "this ad was not created by generative AI." The ad creator is notified of the results by email or via dashboard. At the same time, the emotion engine analyzes the ad creator's emotions, and the content of notifications and reports is optimized according to their emotions.

[1255] The above is a specific embodiment of a generative AI content determination system that combines an emotion engine. This not only enables generative AI to identify content, but also provides highly reliable information that takes into account the user's emotions.

[1256] The processing flow will be explained below.

[1257] Step 1:

[1258] Users upload content from their devices, for example, journalists post new articles and advertising creators upload advertising content.

[1259] Step 2:

[1260] The server receives the content sent by the user. The received data can be in various formats such as text, images, and videos.

[1261] Step 3:

[1262] The content received by the server is preprocessed. During this preprocessing, text data is extracted. If the content is in a specific format such as HTML or PDF, only the text portion is extracted and cleaned up. Unnecessary tags and special characters are removed.

[1263] Step 4:

[1264] The server uses optical character recognition (OCR) technology to extract text from images and videos, for example, subtitles and on-screen text in advertising videos.

[1265] Step 5:

[1266] The server extracts the content's metadata, including information about the author, the date and time of posting, the language used, etc. This information is used for subsequent analysis and reporting.

[1267] Step 6:

[1268] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and stylistic patterns. Specific stylistic tendencies and unnatural phrasing are also detected here.

[1269] Step 7:

[1270] The server inputs the extracted features into a generative AI judgment algorithm, which is a pre-trained machine learning model that determines whether the data was created by generative AI, such as a neural network or decision tree model.

[1271] Step 8:

[1272] Based on the results of the judgment algorithm, the server outputs a judgment result as to whether the content was generated by generative AI, including information on the features used.

[1273] Step 9:

[1274] The server generates a report based on the results of the judgment. This report includes the AI's judgment, its rationale, and metadata, if necessary. For example, the report may conclude, "This article was not created by a generative AI."

[1275] Step 10:

[1276] The server notifies the user of the generated report. Notification methods include sending an email or displaying the report on a dashboard. The user can access the dashboard to check the results.

[1277] Step 11:

[1278] The server uses an emotion engine to recognize the user's emotions. This emotion engine analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions. For example, it analyzes whether the user is feeling stressed.

[1279] Step 12:

[1280] The server adjusts the notification method and report content based on the user's emotions recognized by the emotion engine. If an emotion is recognized, the notification can be made softer or the frequency of notifications can be reduced. For example, if the user is feeling anxious, additional information can be provided to alleviate the anxiety.

[1281] These are the specific processing steps for combining an emotion engine with a generative AI content assessment system, which enables the generative AI to identify content with high accuracy and provides highly reliable information that takes into consideration the user's emotions.

[1282] Example 2

[1283] 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."

[1284] Current generative AI assessment systems not only assess content created by generative AI, but also lack the functionality to appropriately notify users of the assessment results and take their feelings into consideration. As a result, they are unable to reduce user stress and anxiety, which could lead to a decline in system satisfaction.

[1285] 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.

[1286] In this invention, the server includes means for receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result, means for generating a report based on the judgment result, means for notifying the user of the report, and means including an emotion engine for recognizing the user's emotions and adjusting the notification method according to the user's emotions. This enables the generative AI to identify content and to provide notifications that take the user's emotions into consideration.

[1287] The "content receiving means" is a function for receiving content such as articles, advertisements, and reviews from users.

[1288] The "preprocessing means" is a function that removes unnecessary information from the received content and extracts text data.

[1289] The "feature extraction means" is a function that extracts features such as word frequency, sentence structure, and style patterns from preprocessed text data.

[1290] The "generative AI determination algorithm" is an algorithm that uses extracted features to determine whether content was created by generative AI.

[1291] The "judgment result notification means" is a function that generates a report based on the obtained judgment result and notifies the user of it.

[1292] An "emotion engine" is a system that analyzes facial expressions, voice, and text tone to recognize a user's emotions.

[1293] The "means for adjusting the notification method according to emotion" is a function for adjusting the notification method based on the emotion of the user detected by the emotion engine.

[1294] The "metadata extraction means" is a function that extracts metadata from content and uses that information for analysis.

[1295] The present invention is a system that accurately judges content created by a generative AI, and further includes a function to recognize the user's emotions and respond appropriately. Specific embodiments of the system are described below.

[1296] Content reception means

[1297] Users upload content such as articles, advertisements, and reviews from their devices. To do so, users access the system's web interface using their PC or smartphone, select the article file, and click the upload button. Specific examples include a journalist posting a new article, or an advertising creator uploading new advertising content.

[1298] Pretreatment means

[1299] The server receives the content sent by the user and performs preprocessing. This process involves removing unnecessary tags from the received content and extracting clean text data. Specifically, it extracts text from HTML content and extracts text from images and videos using optical character recognition (OCR) technology.

[1300] Feature extraction method

[1301] The server extracts features from the preprocessed text data. This function is achieved by using text analysis tools to extract features such as word frequency, sentence structure, and stylistic patterns. For example, it detects the frequent occurrence of unusual words and grammatical unnaturalness, and stores this information in a database.

[1302] Generative AI Judgment Algorithm

[1303] The server inputs the extracted features into a generative AI determination algorithm. This algorithm uses a pre-trained multilayer neural network model to determine whether the content was created by generative AI. For example, by inputting the features into the algorithm, it can give a score such as "high possibility of being created by generative AI."

[1304] Judgment result notification means

[1305] The server generates a report based on the judgment result and notifies the user. In this process, a report is created that includes the judgment result, the importance of the features, and detailed analysis information, and notifies the user. For example, a PDF report is generated that concludes, "This article was not created by generative AI," and is sent to the user's email address or displayed on the dashboard.

[1306] Emotion Engine

[1307] The server uses an emotion engine to recognize the user's emotions. This engine has the ability to recognize emotions by analyzing the user's facial expressions, voice, and text tone. For example, when a reporter is reviewing a report, it can detect "anxiety" or "stress" from the user's facial expressions and voice.

[1308] Notification method adjustment method according to emotions

[1309] The server adjusts the notification method based on the user's emotions recognized by the emotion engine. If the user is feeling stressed, the server can soften the notification text and adjust the report content to reassure the user. For example, the server can provide additional information such as "If you need more information about the results, please contact support."

[1310] For example, when a media company posts an article, the server receives the article data and extracts the necessary text through preprocessing and analysis. The extracted features are then input into a generative AI judgment algorithm, which ultimately provides a judgment result that "this article was not created by generative AI." The reporter accesses the dashboard to confirm this result. At the same time, the emotion engine analyzes the reporter's facial expressions and voice, and if stress or anxiety is detected, the notification method and report content are adjusted.

[1311] Examples of prompts:

[1312] "How likely is it that this content was created by generative AI?"

[1313] In this way, the present invention is a system that uses generative AI to identify content and adjusts the delivery and notification methods of highly reliable information that takes into account the user's emotions.

[1314] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1315] Step 1: Receiving content

[1316] A user uploads content files such as articles, advertisements, and reviews from a terminal.

[1317] Users access the system's web interface using their PC or smartphone, select the article file, and click the upload button.

[1318] Input: Various content files (text files, HTML files, image files, etc.)

[1319] Output: Content file saved on the server

[1320] Step 2: Preprocessing the data

[1321] The server receives the content sent by the user and performs pre-processing.

[1322] The server removes unnecessary tags from the received HTML file and extracts clean text data.

[1323] For image files, the server uses optical character recognition (OCR) technology to extract the text.

[1324] Input: Saved content file

[1325] Output: Cleaned text data

[1326] Step 3: Feature extraction

[1327] The server extracts features from the preprocessed text data.

[1328] The server uses text analysis tools to identify features such as word frequency, sentence structure, and stylistic patterns.

[1329] Input: Preprocessed text data

[1330] Output: Extracted features (word frequency, sentence structure, stylistic patterns, etc.)

[1331] Step 4: Run the generative AI decision algorithm

[1332] The server inputs the extracted features into the generative AI judgment algorithm.

[1333] The server uses a pre-trained multi-layer neural network model to determine whether the content was created by generative AI.

[1334] Input: extracted features

[1335] Output: Judgment results (score and evaluation) by the generating AI

[1336] Step 5: Output of judgment results

[1337] The server determines whether the content was generated by a generation AI based on the results of the generation AI determination algorithm.

[1338] The server compiles the results of the assessment and the reasons for it in a report.

[1339] Input: Results of the generative AI decision algorithm

[1340] Output: Report containing the judgement results

[1341] Step 6: Generate reports

[1342] The server generates a detailed report based on the results of the assessment.

[1343] The report includes details of the judgment results, the reasons for the judgment, and the results of the feature analysis.

[1344] Input: Judgment result and its reasoning

[1345] Output: Detailed report

[1346] Step 7: Notification of results

[1347] The server notifies the user of the generated report.

[1348] The server sends the report to the user's email address or displays it in a dashboard that the user accesses.

[1349] Input: Generated report

[1350] Output: Email notification or dashboard notification

[1351] Step 8: Use the Emotion Engine

[1352] The server uses an emotion engine to recognize the user's emotions.

[1353] The server analyzes the user's facial expressions, voice, text tone, etc. to recognize emotions.

[1354] Input: User facial expressions, tone of voice, and text

[1355] Output: User emotion recognition results (e.g., anxiety, stress, joy, etc.)

[1356] Step 9: Respond emotionally

[1357] The server adjusts the notification method based on the user's emotion recognized by the emotion engine.

[1358] The server changes the notification wording to softer expressions depending on the emotion and provides additional information as needed.

[1359] Input: User emotion recognition results

[1360] Output: Adjusted notification wording and additional information

[1361] (Application example 2)

[1362] 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."

[1363] Recently, the amount of content created by generative AI has been increasing, and there is a demand for systems that can accurately distinguish between them. However, it is important not only to simply distinguish between AI-generated content, but also to provide optimal information by taking into consideration the emotions of the user viewing the content. However, current systems have insufficient emotion recognition capabilities, which means they are unable to improve the quality of the user experience.

[1364] 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 receiving content from a user, means for preprocessing the received content to extract text data, means for extracting features from the preprocessed text data, means for inputting the extracted features into a generation AI judgment algorithm to obtain a judgment result, means for recognizing the user's emotions using the smartphone's camera and microphone, means for optimizing the content to be displayed based on the user's emotions, and means for notifying the user of the judgment result. This not only enables the generation AI to accurately distinguish content, but also enables the provision of appropriate content according to the user's emotions.

[1365] "Generative AI" is a system that uses artificial intelligence technology to automatically generate creative content in the same way that humans do.

[1366] "Content" is a collective term for digital information such as text, images, video, and audio that is created and shared by users.

[1367] "Means for receiving" refers to a mechanism, device, or program for receiving content sent from a user at a server or terminal.

[1368] The "means for preprocessing and extracting text data" is a function for unifying the data format of received content and performing processing to extract the necessary text data.

[1369] "Means for extracting features" refers to a method or device for identifying and quantifying useful patterns and attributes from preprocessed text data.

[1370] A "generative AI determination algorithm" is an algorithm that determines whether the input features were generated by a generative AI based on specific rules or models.

[1371] "Means for recognizing emotions" refers to devices or software that analyze the user's voice and facial expressions and estimate their emotional state.

[1372] The "means for optimizing the content to be displayed" is a function for adjusting the content and order of the content to be displayed according to the emotional state of the user.

[1373] System configuration

[1374] This invention is a system that recognizes user emotions and accurately judges content created by generative AI. The system includes the following main components:

[1375] 1. Content Reception Method

[1376] Users upload content such as articles, videos, and audio from their devices. For example, this is how users post news articles.

[1377] 2. Pretreatment Methods

[1378] The server receives the content submitted by the user and pre-processes it to extract text data. For example, if the content is in HTML format, the text portion is extracted and cleaned up. If the content contains images or videos, optical character recognition (OCR) techniques are used to extract text from them.

[1379] 3. Feature Extraction Method

[1380] The server extracts features from the preprocessed text data, including word frequency, sentence structure, and specific patterns of writing style. This process extracts specific patterns and unnatural expressions within the text.

[1381] 4. Generative AI Judgment Algorithm

[1382] The server inputs the extracted features into a generative AI determination algorithm. A pre-trained machine learning model is used to determine whether the content was created by generative AI. For example, a machine learning model such as a multilayer neural network or decision tree is used.

[1383] 5. Emotion recognition means

[1384] The server uses the smartphone's camera and microphone to recognize the user's emotions. This includes technology that analyzes the user's facial expressions and voice to recognize emotions. It uses facial recognition APIs such as Google's MediaPipe and Azure's Emotion API.

[1385] 6. Feed optimization methods

[1386] The server optimizes the content feed displayed to the user based on the data obtained from the emotion recognition means. For example, if the user is feeling stressed, it will prioritize displaying relaxing content.

[1387] 7. Judgment result notification means

[1388] The server notifies the user of the results. Notification methods include sending an email or displaying the results on the dashboard. The user can access the dashboard to check the results.

[1389] Examples and prompts

[1390] Below is a concrete example of how the system actually works.

[1391] Example: Checking news articles

[1392] Suppose a user reads a new news article on their smartphone. First, the smartphone's microphone records the audio data, and the facial expression camera captures the user's facial expressions. These data are sent to the server, where the emotion recognition means recognizes the emotions. An example of a prompt sentence in this case is as follows:

[1393] Example prompt sentence:

[1394] Determine if the following article was created with generative AI:

[1395] "We bring you the latest economic news about..." (article content)

[1396] The server analyzes the user's emotional state based on the emotion recognition results. Then, a preprocessing means extracts the article's text data and generates features. The generated features are input into a generative AI judgment algorithm to determine whether the article was created by generative AI. Finally, the judgment result is notified to the user, and the feed is optimized based on the user's emotions.

[1397] In this way, users can verify the authenticity of an article and at the same time be provided with appropriate information that matches their emotions at the time.

[1398] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1399] Step 1:

[1400] A user inputs content using a terminal. For example, a user posts a news article on a smartphone. The input content is sent to a server via an application on the smartphone. This is the content receiving means.

[1401] Step 2:

[1402] The server receives content sent by the user using the content receiving means. It then analyzes the content using the preprocessing means and extracts text data. For example, it extracts text from HTML content and performs cleanup processing. It also extracts text from images and videos using OCR technology as needed. The input is the content, and the output is organized text data.

[1403] Step 3:

[1404] The server extracts features from the preprocessed text data. Features include word frequency, sentence structure, and stylistic patterns. This allows specific patterns and unnatural expressions in the text to be quantified. The input is the preprocessed text data, and the output is feature data.

[1405] Step 4:

[1406] The server uses a generative AI judgment algorithm to determine whether the content was created by generative AI based on the extracted features. For example, a pre-trained machine learning model such as a multilayer neural network or decision tree is used. The input is the feature data, and the output is the judgment result.

[1407] Step 5:

[1408] The server recognizes the user's emotions using the smartphone's camera and microphone. The camera captures the user's facial expressions and analyzes them using a facial expression recognition API (such as Google's MediaPipe). The server also collects audio data using the microphone and performs audio emotion analysis. The input is the user's facial expression data and audio data, and the output is the emotion recognition results.

[1409] Step 6:

[1410] The server optimizes the content feed to be displayed to the user based on the emotion recognition results. For example, if the user is feeling stressed, it will prioritize displaying relaxing content. The input is the emotion recognition results, and the output is the optimized content feed.

[1411] Step 7:

[1412] The server notifies the user of the results of the generated AI's judgment. Notification methods include displaying the results on the dashboard or sending an email. The user can access the dashboard and check the contents. The input is the judgment result and feed optimization result, and the output is a user notification.

[1413] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice 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 voice data.

[1414] 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.

[1415] 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 robot 414.

[1416] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1417] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1418] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1419] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1420] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1421] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1422] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1423] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1424] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1425] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1426] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1427] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1428] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1429] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1430] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1431] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1432] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1433] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1434] The following is further disclosed regarding the above embodiment.

[1435] (Claim 1)

[1436] A system for determining whether content was created by generative AI,

[1437] means for receiving content from a user;

[1438] means for preprocessing the received content to extract text data;

[1439] means for extracting features from the preprocessed text data;

[1440] A means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result;

[1441] means for notifying a user of the determination result;

[1442] A system including:

[1443] (Claim 2)

[1444] 10. The system of claim 1, further comprising means for extracting metadata of the content.

[1445] (Claim 3)

[1446] 10. The system of claim 1, further comprising means for generating a report based on the determination result.

[1447] "Example 1"

[1448] (Claim 1)

[1449] A system for determining whether content was created by generative AI,

[1450] means for receiving content from a user;

[1451] means for preprocessing the received content to extract text data;

[1452] means for extracting features from the preprocessed text data;

[1453] A means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result;

[1454] A means for generating a report based on the determination results;

[1455] means for notifying a user of the determination result;

[1456] A system including:

[1457] (Claim 2)

[1458] 10. The system of claim 1, further comprising means for extracting metadata of the content.

[1459] (Claim 3)

[1460] 10. The system of claim 1, further comprising means for using character recognition techniques to extract text from the image or video.

[1461] "Application Example 1"

[1462] (Claim 1)

[1463] A system for determining whether content was created by generative AI,

[1464] means for receiving content from a user;

[1465] means for preprocessing the received content to extract text data;

[1466] means for extracting features from the preprocessed text data;

[1467] A means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result;

[1468] means for notifying a user of the determination result;

[1469] A means for capturing images of advertisements using a camera on a specific device (smart glasses) and extracting text in real time;

[1470] a means for visually displaying the determination results on a device display in real time;

[1471] A system including:

[1472] (Claim 2)

[1473] 10. The system of claim 1, further comprising means for extracting metadata of the content.

[1474] (Claim 3)

[1475] 10. The system of claim 1, further comprising means for generating a report based on the determination result.

[1476] "Example 2: Combining Emotion Engines"

[1477] (Claim 1)

[1478] A system for determining whether content was created by generative AI,

[1479] means for receiving content from a user;

[1480] means for preprocessing the received content to extract text data;

[1481] means for extracting features from the preprocessed text data;

[1482] A means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result;

[1483] A means for generating a report based on the determination results;

[1484] a means for notifying a user of the report;

[1485] means for adjusting a notification method according to the user's emotion, the emotion engine including the emotion engine for recognizing the user's emotion;

[1486] A system including:

[1487] (Claim 2)

[1488] 10. The system of claim 1, further comprising means for extracting metadata of the content.

[1489] (Claim 3)

[1490] 10. The system of claim 1, wherein the emotion engine further comprises means for analyzing a user's facial expression, voice, and text tone to recognize the user's emotion.

[1491] "Application example 2 when combining emotion engines"

[1492] (Claim 1)

[1493] A system for determining whether content was created by generative AI,

[1494] means for receiving content from a user;

[1495] means for preprocessing the received content to extract text data;

[1496] means for extracting features from the preprocessed text data;

[1497] A means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result;

[1498] A means for recognizing user emotions using a smartphone camera or microphone;

[1499] A means for optimizing content to be displayed based on user emotions;

[1500] means for notifying a user of the determination result;

[1501] A system including:

[1502] (Claim 2)

[1503] 10. The system of claim 1, further comprising means for extracting metadata of the content.

[1504] (Claim 3)

[1505] 10. The system of claim 1, further comprising means for generating a report based on the determination result. [Explanation of symbols]

[1506] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A system for determining whether content was created by generative AI, means for receiving content from a user; means for preprocessing the received content to extract text data; means for extracting features from the preprocessed text data; A means for inputting the extracted features into a generative AI judgment algorithm to obtain a judgment result; means for notifying a user of the determination result; A system including:

2. The system of claim 1 further comprising means for extracting metadata of the content.

3. The system of claim 1 further comprising means for generating a report based on the determination result.

Images