System

Abstract

A system is provided.SOLUTION: A system comprising: means for automatically filtering generated content; means for assessing a risk of piracy of the generated content; means for blocking the generated content if the generated content contains inappropriate content; and means for notifying a user of the generated content.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] Advances in generative AI have enabled the automatic generation of a large amount of content, but this has several serious issues. Specifically, the generated content may contain inappropriate content and may pose a risk of copyright infringement. These issues undermine the trust users and companies place in using generative AI and also pose legal risks. The present invention aims to provide a system that solves these issues and ensures the appropriateness and legal safety of generated content. [Means for solving the problem]

[0005] The present invention provides a means for automatically filtering generated content and assessing the risk of copyright infringement. Specifically, it includes a means for analyzing generated content using natural language processing and blocking it if it contains inappropriate content. It also includes a means for comparing the similarity of generated content with an existing database to assess the risk of copyright infringement. Furthermore, it provides a means for notifying users of the results of the generated content, allowing users to use generated content with peace of mind.

[0006] "Means for automatically filtering generated content" refers to a function that analyzes content generated by the generative AI and automatically detects whether it contains inappropriate content or expressions.

[0007] "Means for assessing the risk of copyright infringement of generated content" refers to a function that checks whether generated content is similar to existing works and determines the possibility of copyright infringement.

[0008] "Means for blocking generated content if it contains inappropriate content" refers to a function that automatically removes content in which inappropriate content is detected before providing it to users.

[0009] "Means for notifying users about generated content" refers to a function that communicates the results of filtering and moderation to users and notifies them whether the generated content is appropriate or inappropriate and has been blocked.

[0010] "Natural language processing" is a technology that uses computers to understand and process human language, and is used to analyze text data and understand its meaning.

[0011] "Means for matching similarities with existing databases" refers to a function that compares generated content with content in existing databases to determine similarities that may constitute potential copyright infringement.

[0012] Based on these definitions, a system is constructed to improve the appropriateness and legal safety of generated content. [Brief explanation of the drawings]

[0013] [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

[0014] 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.

[0015] First, the terms used in the following description will be explained.

[0016] 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).

[0017] 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.

[0018] 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.

[0019] 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.

[0020] 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."

[0021] [First embodiment]

[0022] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0023] 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.

[0024] 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).

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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.

[0029] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] 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.

[0031] 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.

[0032] 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.

[0033] 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."

[0034] The present invention relates to a system for automatically filtering and moderating content generated by generative AI to reduce the risk of inappropriate content and copyright infringement. The system provides a means for analyzing the generated content and detecting the risk of inappropriate content and copyright infringement.

[0035] System Overview

[0036] The system includes the following main functions:

[0037] 1. Automatically filtering generated content

[0038] 2. Means of assessing the risk of copyright infringement of generated content

[0039] 3. How to block generated content if it contains inappropriate content

[0040] 4. Means of notifying users about generated content

[0041] Program processing

[0042] The program of this system performs processing in the following procedure.

[0043] 1. Submitting a content generation request

[0044] The user uses a device to send a content generation request to the generation AI, which includes the type of content they want to generate and guidelines.

[0045] 2. Receiving and forwarding requests

[0046] The server receives the request from the user and forwards the request to the generation AI.

[0047] 3. Content generation using generative AI

[0048] The generation AI generates the specified content based on the request.

[0049] 4. Receiving Generated Content

[0050] The server receives the generated content from the generation AI.

[0051] 5. Start filtering and moderation

[0052] Server-generated content is automatically filtered and moderated, using natural language processing (NLP) to analyze text and detect inappropriate language, and it also compares content similarity with existing databases to assess the risk of copyright infringement.

[0053] 6. Blocking inappropriate content

[0054] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[0055] 7. Notice to Users

[0056] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0057] Specific examples

[0058] 1. User Request Submission

[0059] A user submits a request to "generate an adventure story for children."

[0060] 2. Receiving and forwarding requests

[0061] The server receives the request and forwards it to the generation AI.

[0062] 3. Story Generation Using Generative AI

[0063] The AI ​​generates an adventure story in which a prince and princess encounter mysterious creatures in the forest.

[0064] 4. Receiving Generated Content

[0065] The server receives the generated story.

[0066] 5. Filtering and moderation processes

[0067] The server analyzes the story text and checks for inappropriate language and similarities to existing works. For example, if a story contains a scene in which the prince indiscriminately defeats monsters, it will automatically detect it as inappropriate.

[0068] 6. Blocking inappropriate content

[0069] The server detects inappropriate content and blocks the story.

[0070] 7. Notifying Users and Providing Results

[0071] The server sends a notification to the user saying "Generation blocked due to inappropriate content."

[0072] As a result, the present invention provides an environment in which users can use generative AI with peace of mind, improving the quality of generated content and reducing legal risks.

[0073] The processing flow will be explained below.

[0074] Program processing flow

[0075] Step 1:

[0076] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[0077] Step 2:

[0078] The server receives the user's request, which includes details about the content, and prepares to parse it and forward it to the generation AI in the appropriate format.

[0079] Step 3:

[0080] The server forwards the content generation request to the generation AI, which starts the process of the generation AI generating content based on the specified conditions.

[0081] Step 4:

[0082] Generative AI generates content based on user requests, such as "A story about a prince and princess having an adventure in the forest."

[0083] Step 5:

[0084] The server receives the generated content from the generation AI. This content is raw, as it has not yet been filtered or moderated.

[0085] Step 6:

[0086] The server then begins the filtering and moderation process on the generated content, first using natural language processing (NLP) to analyze the text and check for inappropriate language or expressions.

[0087] Step 7:

[0088] The server then compares the content similarity with existing databases to assess the risk of copyright infringement of the generated content. If similar content is found, the content is deemed to be at risk of copyright infringement.

[0089] Step 8:

[0090] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters would be deemed inappropriate.

[0091] Step 9:

[0092] The server generates a result indicating whether the content is appropriate or not and notifies the user. If the content is appropriate, it is served to the user as is, otherwise an error message is sent.

[0093] Step 10:

[0094] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they use it, and if it is inappropriate, they send a request again.

[0095] In this way, the system automatically checks the quality and legal risks of generated content, providing a safe and secure environment for users.

[0096] Example 1

[0097] 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."

[0098] Since content generated using generative AI may contain inappropriate content or may infringe copyrights, there is no environment in place for users to use it safely. For this reason, there is a need for a system that can automatically filter generated content and provide only appropriate content.

[0099] 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.

[0100] In this invention, the server includes: a means for a user to send a content generation request to the generation AI using a terminal; a means for the server to receive the request from the user and forward it to the generation AI; a means for the generation AI to generate content based on the request; a means for the server to receive the generated content from the generation AI; a means for the server to automatically filter and moderate the generated content; a means for the server to block the generated content if the server determines that the generated content contains inappropriate content or poses a risk of copyright infringement; and a means for the server to notify the user of the result of whether the generated content is appropriate or has been blocked. This makes it possible to improve the quality of generated content and reduce legal risks.

[0101] "User" means an individual or entity that uses a device to send a content generation request to a generation AI.

[0102] A "terminal" is an electronic device used by a user to send a content generation request to the generation AI, including a computer, smartphone, tablet, etc.

[0103] "Generative AI" is an artificial intelligence system that automatically generates content based on a given request.

[0104] A "server" is a computer system that receives requests from users via the Internet, forwards them to the generation AI, and receives and processes the generated content.

[0105] A "content generation request" is an instruction sent by a user to a generation AI to generate specific content, including the content and guidelines.

[0106] "Filtering" is the process of automatically detecting and filtering out inappropriate content from generated content.

[0107] "Moderation" is the process of evaluating and managing generated content to detect inappropriate content or the risk of copyright infringement.

[0108] "Notification" is the act of the server informing the user whether the generated content is appropriate or has been blocked because it is inappropriate.

[0109] "Natural Language Processing (NLP)" refers to the technology of analyzing, understanding, and appropriately processing human language. It is used for text analysis of generated content.

[0110] "Copyright infringement" is the act of using someone else's copyright without permission, which carries legal risks.

[0111] "Blocking" is the act of restricting access to prevent users from being provided with inappropriate content or generated content that poses a risk of copyright infringement.

[0112] This invention relates to a system that automatically filters and moderates content generated by generative AI, aiming to improve the quality of generated content and reduce legal risks. This system is composed of elements such as users, terminals, servers, and generative AI, and by clarifying the roles of each, it achieves efficient and safe content generation.

[0113] System Configuration

[0114] User

[0115] Users send content generation requests to the generative AI using a dedicated device, which can be a computer, smartphone, tablet, or other device, and access the generative AI model using a web browser or dedicated application.

[0116] Terminal

[0117] The terminal is a device through which users can send content generation requests to the generation AI, and acts as a user interface. A request form is displayed here, and users can enter the type of content they want to generate and guidelines.

[0118] server

[0119] The server receives user requests and forwards them appropriately to the AI ​​generator. It also receives the generated content returned by the AI ​​generator and automatically filters and moderates it. Specific software used in this process includes the natural language processing (NLP) libraries SpaCy and NLTK.

[0120] Generation AI

[0121] Generative AI is an artificial intelligence system that generates text content based on user requests, including language models such as GPT-4®.

[0122] Specific examples

[0123] 1. User Request Submission

[0124] A user sends a request from their device to "generate an adventure story for children." Specifically, they enter "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or pose a risk of copyright infringement" into the web form and click the submit button.

[0125] 2. Receiving and forwarding requests

[0126] The server receives the HTTP request, analyzes its contents, and converts it into JSON format for transmission to the API of the AI ​​generation. At this time, the request contents are transmitted to the endpoint of the AI ​​generation.

[0127] 3. Story Generation Using Generative AI

[0128] Based on the request, the AI ​​generates an "adventure story about a prince and princess meeting mysterious creatures in the forest." The generated content is sent back to the server in JSON format.

[0129] 4. Receiving and Processing Generated Content

[0130] The server receives the generated stories and stores them in an internal data structure, then uses an NLP library (e.g., SpaCy or NLTK) to analyze the text and check for profanity and similarity to existing works.

[0131] 5. Blocking inappropriate content

[0132] If the server determines that the content contains inappropriate content based on the analysis results, it will block the content. For example, if the content contains a scene in which the prince indiscriminately kills monsters, the filtering function will automatically detect it as inappropriate.

[0133] 6. Notice to Users

[0134] The server notifies the user whether the generated content is appropriate or has been blocked because it is inappropriate. Specifically, the user is shown a message saying, "The generation has been blocked because it contains inappropriate content."

[0135] Prompt Sentence Examples

[0136] Use the following prompt to input the generative AI model:

[0137] Create an adventure story for children. Please be careful not to include violent scenes or copyright infringement in the content.

[0138] This system allows users to use generative AI with confidence, reducing legal risks while producing high-quality content.

[0139] keyword

[0140] Generative AI model, prompt sentence

[0141] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0142] Step 1:

[0143] The user uses their device to send a content generation request to the generation AI. This request includes the type of content they want to generate and guidelines. Specifically, the user enters the following into a web form: "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or risk copyright infringement." and clicks the submit button. This form input is sent to the server as a request.

[0144] input:

[0145] User input into a web form (content generation request)

[0146] output:

[0147] HTTP request to the server

[0148] Step 2:

[0149] The server receives the request sent by the user. It analyzes the received request and converts it into a format suitable for the API of the generation AI (for example, JSON format). It then forwards the request to the endpoint of the generation AI. Specifically, the server checks the contents of the request, converts the format if necessary, and sends it to the API of the generation AI.

[0150] input:

[0151] HTTP request from the user

[0152] output:

[0153] The converted request to the generating AI (JSON format)

[0154] Step 3:

[0155] The generation AI generates content based on the received request. The generation AI uses an internal algorithm (e.g., GPT-4) to automatically generate text based on the input prompt. The generated content is sent back to the server in JSON format.

[0156] input:

[0157] The transformed request to the generative AI

[0158] output:

[0159] Generated content (JSON format)

[0160] Step 4:

[0161] The server receives the generated content from the generation AI, stores it in an internal data structure, and prepares it for the next filtering step. Specifically, the server parses the JSON-formatted data and extracts the text portion of the content.

[0162] input:

[0163] Generated content (JSON format)

[0164] output:

[0165] Storing text in internal data structures

[0166] Step 5:

[0167] The server automatically filters and moderates the generated content. During this process, it uses natural language processing (NLP) libraries (such as SpaCy or NLTK) to analyze the text and check for inappropriate content and similarities. Specifically, text analysis detects violent scenes and inappropriate content, and compares the generated content's similarities with existing databases. For example, if a video contains a scene in which a prince indiscriminately defeats monsters, it will be detected as violent.

[0168] input:

[0169] Generated content text

[0170] output:

[0171] Filtering and Moderation Results

[0172] Step 6:

[0173] If the server determines that the content is inappropriate or poses a risk of copyright infringement as a result of filtering and moderation, it will block the content. Specifically, if the filtering result is negative, the generated content will not be returned to the user and an error message will be prepared.

[0174] input:

[0175] Filtering and Moderation Results

[0176] output:

[0177] Generate appropriate content or error messages

[0178] Step 7:

[0179] The server notifies the user whether the generated content is appropriate or inappropriate and has been blocked. The server returns the generated content or an error message to the user's device as an HTTP response. Specifically, a message stating "Generation has been blocked because it contains inappropriate content" is displayed.

[0180] input:

[0181] Appropriate content or error messages

[0182] output:

[0183] HTTP response to the user

[0184] ---

[0185] The above is the specific flow of the program processing of this system. This allows users to use the generative AI with peace of mind, improves the quality of generated content, and reduces legal risks.

[0186] (Application example 1)

[0187] 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."

[0188] If the content generated by generative AI contains inappropriate content or poses a risk of copyright infringement, it is difficult to provide a safe environment for users. Furthermore, there is a lack of effective means to check for these risks before the generated content is distributed, raising concerns about legal risks and a decline in quality. In particular, in the case of educational content, the inclusion of inappropriate content not only reduces the educational effectiveness but also reduces the credibility of the content.

[0189] 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.

[0190] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying users of the generated content, means for checking the generated content before distribution, means for analyzing the audio and subtitles of the generated content using natural language processing, and means for evaluating the generated content based on guidelines specialized for educational content. This allows the generated content to be used and distributed safely without containing inappropriate content or the risk of copyright infringement.

[0191] "Generated content" refers to digital content generated by a generative AI in response to a user request.

[0192] "Filtering" refers to the process of automatically detecting and removing inappropriate language or content contained in generated content.

[0193] "Risk of copyright infringement" refers to situations in which generated content may infringe the rights of existing copyrighted works.

[0194] "Inappropriate content" refers to any expression or information that may violate user guidelines or the law.

[0195] "Notification to users" refers to a means of notifying users of the evaluation results and block information of generated content.

[0196] The "means for checking before distribution" refers to a process for confirming the appropriateness of generated content before distributing the content to a user or a third party.

[0197] "Natural language processing" refers to the technology that enables computers to understand, generate, and interpret human language.

[0198] "Means for analyzing audio and subtitles" refers to technology for automatically analyzing the audio data and subtitle data contained in the generated content and evaluating its content.

[0199] "Educational Content Specific Guidelines" means specific standards or rules that apply to generated content used for educational purposes.

[0200] This invention is a system for automatically filtering and moderating content generated by generative AI, with the aim of reducing the risk of inappropriate content and copyright infringement, particularly in the generation of educational content, by eliminating inappropriate content and providing reliable, high-quality content.

[0201] System Configuration

[0202] The system consists of the following main components:

[0203] 1. Server: Receives requests for generated content and forwards them to the generation AI. It then receives the generated content and performs filtering and moderation. The hardware used is, for example, an AWS (registered trademark) EC2 instance.

[0204] 2. Generative AI: Generates content based on user requests. For example, OpenAI's GPT-4 is used.

[0205] 3. Natural language processing tools: Analyze the audio and subtitles of generated content to assess the risk of inappropriate content and copyright infringement. For example, Google® Cloud Natural Language API is used.

[0206] 4. User device: Sends requests and receives generated content. Uses a smartphone (iOS or ANDROID (registered trademark) device).

[0207] Data processing and calculation

[0208] Receiving and forwarding requests: The server forwards requests for generated content received from the user device to the generation AI. The request includes the type of content to be generated and guidelines.

[0209] Content Generation: Generative AI generates content based on given guidelines, for example, given a prompt to generate an "educational video learning about the minerals of the Earth."

[0210] Filtering and moderation: The server automatically filters and moderates generated content, using natural language processing (Google Cloud Natural Language API) to analyze text and audio data and assess risk of inappropriate content and copyright infringement.

[0211] Notification to users: Based on the evaluation results, users will be notified whether the generated content is appropriate. If it is deemed inappropriate, users will be notified along with the reason.

[0212] Specific examples

[0213] A user might submit a request to generate educational video content, for example, "Please generate an educational video about the minerals of the Earth. Please do not include any inappropriate language or copyright infringing content."

[0214] The server forwards this request to the generation AI, which generates an educational video according to the specified guidelines. The server then receives the generated video and analyzes the video's audio and subtitles using natural language processing tools. If the analysis finds that the content does not comply with the educational guidelines or poses a risk of copyright infringement, the video is blocked. Finally, the server notifies the user of the appropriateness of the generated content.

[0215] In this way, users can confidently create and distribute high-quality, reliable educational content.

[0216] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0217] Step 1:

[0218] The user uses their device to send a request to the AI. The request includes the type of content they want to generate and guidelines. For example, they might enter a prompt like, "Generate an educational video that teaches about the minerals on Earth. Please do not include any inappropriate language or copyright infringement." This request is then sent to the server.

[0219] Step 2:

[0220] The server parses the request received from the user and converts it into an appropriate format. The server then forwards the request to the generation AI, which sends the request data to a specific API endpoint and applies prompts so that the generation AI can understand the content.

[0221] Step 3:

[0222] The generation AI generates content based on the request. The generation AI receives prompt text as input and outputs generated content such as an educational video script or narration. Each generation AI has different parameters and settings, so processing is done in the appropriate data format.

[0223] Step 4:

[0224] The server receives the generated content from the generation AI, records its contents, verifies the data format of the generated content (video script and narration), and converts it to fit the next processing step.

[0225] Step 5:

[0226] The server analyzes the generated content using natural language processing (NLP) tools. Audio and subtitle data is converted into text format and analyzed through a natural language processing engine (e.g., Google Cloud Natural Language API). The results of this analysis are used to assess the risk of inappropriate content and copyright infringement.

[0227] Step 6:

[0228] The server filters the generated content based on the analysis results. If inappropriate content or a risk of copyright infringement is detected, the server blocks the content. If necessary, the server modifies part of the generated content or requests the AI ​​to regenerate it.

[0229] Step 7:

[0230] The server notifies the user of the evaluation results of the generated content. If the content is appropriate, the server notifies the user and provides a download link and notification that the content is ready for distribution. If the content contains inappropriate content, the server notifies the user, along with the reason for the inappropriate content.

[0231] In this way, through a series of processing steps, a system is realized that ensures the quality of user-generated content and provides an environment in which it can be distributed safely.

[0232] 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.

[0233] This invention relates to a system that provides more appropriate content by combining a system that reduces the risk of inappropriate content and copyright infringement in content generated by generative AI with an emotion engine that recognizes user emotions. By taking user emotions into consideration, this system aims to further improve the quality of generated content and enhance the user experience.

[0234] System Overview

[0235] The system includes the following main functions:

[0236] 1. Automatically filtering generated content

[0237] 2. Means of assessing the risk of copyright infringement of generated content

[0238] 3. How to block generated content if it contains inappropriate content

[0239] 4. Means of notifying users about generated content

[0240] 5. Emotion engine that recognizes user emotions

[0241] 6. The ability to adjust generated content based on user sentiment

[0242] 7. Means for suspending provision of generated content based on emotion recognition results

[0243] Program processing

[0244] The program of this system performs processing in the following procedure.

[0245] 1. Activating the Emotional Engine

[0246] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time. This analysis recognizes the user's emotions.

[0247] 2. Submitting a content generation request

[0248] The user sends a content generation request to the AI ​​via their device, which includes the type of content they want to generate and guidelines.

[0249] 3. Receiving and forwarding requests

[0250] The server receives the user's request and forwards it to the generation AI.

[0251] 4. Content generation using generative AI

[0252] The generation AI generates content that meets the specified conditions based on the request.

[0253] 5. Receiving Generated Content

[0254] The server receives the generated content from the generation AI.

[0255] 6. Start filtering and moderation

[0256] Automatic filtering and moderation of server-generated content, using natural language processing (NLP) to analyze text and check for inappropriate language and copyright infringement.

[0257] 7. Emotion-aware content adjustment

[0258] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state and adjusts the content as necessary. For example, if the user is feeling stressed, the server may provide a relaxing story.

[0259] 8. Blocking inappropriate content

[0260] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[0261] 9. Notice to Users

[0262] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0263] Specific examples

[0264] 1. User Emotion Recognition

[0265] While the user is viewing content on their mobile device, the emotion engine analyzes the user's facial expression data in real time and recognizes their "enjoyment" state.

[0266] 2. Sending a content generation request

[0267] A user submits a request to "generate an adventure story for children."

[0268] 3. Adventure Story Generation Using Generative AI

[0269] The AI ​​generates an "adventure story," which includes dangerous adventure scenes.

[0270] 4. Filtering and Moderation

[0271] The server performs filtering and moderation, determining that dangerous scenes are inappropriate and blocking them.

[0272] 5. Adjustment based on emotion recognition results

[0273] In order to keep the user in a "fun" state, the emotion engine adjusts to prioritize the generation of new stories based on "fun adventures."

[0274] 6. Delivery of Final Content

[0275] The server generates a tailored and appropriate adventure story and provides it to the user.

[0276] In this way, the system can provide appropriate content while taking into account the user's emotions, improving the overall user experience.

[0277] The processing flow will be explained below.

[0278] Specific processing flow of the program

[0279] Step 1:

[0280] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time, thereby recognizing the user's emotional state.

[0281] Step 2:

[0282] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[0283] Step 3:

[0284] The server receives the user's request, analyzes the request content, and prepares to forward it to the generation AI in an appropriate format.

[0285] Step 4:

[0286] The server forwards a content generation request to the generation AI, and the generation AI starts generating content based on the specified conditions.

[0287] Step 5:

[0288] The generation AI generates the specified content based on the request, for example, "A story about a prince and princess having an adventure in the forest."

[0289] Step 6:

[0290] The server receives the generated content from the generative AI, which has not yet been filtered or moderated.

[0291] Step 7:

[0292] The server analyzes the generated content using natural language processing (NLP) to check for inappropriate words and expressions.

[0293] Step 8:

[0294] The server evaluates the risk of copyright infringement of generated content by matching the content similarity with existing databases. If similar content is found, the content is determined to be at risk of copyright infringement.

[0295] Step 9:

[0296] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state. For example, if the user is feeling anxious, the content can be adjusted to change it into a relaxing story.

[0297] Step 10:

[0298] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters is deemed inappropriate and blocked.

[0299] Step 11:

[0300] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0301] Step 12:

[0302] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they can use it, and if it is inappropriate, they are given the option to send the request again or receive the adjusted content.

[0303] Specific examples

[0304] 1. Step 1-6: The user sends a request to "generate an adventure story for children," and the server receives the request and forwards it to the generation AI. The generation AI generates a story about a prince and princess having an adventure in the forest, and the server receives it.

[0305] 2. Step 7: The server analyzes the generated story using natural language processing, for example, to check whether it contains any "dangerous scenes."

[0306] 3. Step 8: The server checks the generated content against an existing database to assess the risk of copyright infringement.

[0307] 4. Step 9: The emotion engine recognizes the user's "anxiety" state, and the server adjusts the story content to a "relaxing adventure story" based on this result.

[0308] 5. Steps 10-12: The server determines that the final adjusted story is appropriate and notifies the user, who then accepts and views the story.

[0309] In this way, the system automatically checks the quality and legal risks of generated content, and further adjusts the content based on user sentiment, providing an environment where users can use it with peace of mind.

[0310] Example 2

[0311] 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."

[0312] The content generated by modern generative AI models may contain inappropriate content or risk copyright infringement. Furthermore, if the generated content does not match the user's emotional state, it can negatively impact the user experience. To address this issue, it is necessary to ensure the appropriateness of the content and take the user's emotional state into consideration.

[0313] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying the user about the generated content, means for recognizing the user's emotion, means for adjusting the content of the generated content based on the user's emotion, and means for suspending provision of the generated content based on the result of the user's emotion recognition. This makes it possible to provide content that matches the user's emotional state while ensuring the appropriateness of the generated content.

[0314] "Generated content" refers to information such as text, images, and video that is automatically generated by a generative AI model based on a user request.

[0315] "Filtering measures" are technologies or algorithms used to analyze generated content and determine whether it contains inappropriate content or is at risk of copyright infringement.

[0316] "Copyright infringement risk assessment means" means any technology or algorithm that compares generated content with existing databases or copyrighted material to assess the risk of copyright infringement.

[0317] "Blocking measures" are technologies or methods that remove generated content before it is provided to users if the content contains inappropriate content or poses a risk of copyright infringement.

[0318] A "notification mechanism" is a mechanism for informing users of the outcome of generated content, whether it is appropriate or blocked for being inappropriate.

[0319] "Emotion recognition means" refers to technology or algorithms that analyze a user's facial expressions, voice, and input data, and recognize the user's emotions in real time based on the results.

[0320] The "emotion-based content adjustment means" refers to a technology or algorithm for changing the content of generated content to match the emotional state of the user based on the result of the user's emotion recognition.

[0321] The "provision interruption means" is a technique or method for interrupting the provision of generated content midway when the user's emotion recognition result satisfies certain conditions.

[0322] MODE FOR CARRYING OUT THE INVENTION

[0323] This invention is a system that reduces the risk of inappropriate content and copyright infringement in content generated using a generative AI model, and combines it with an emotion engine that recognizes user emotions to provide more appropriate content. This system aims to further improve the quality of generated content and enhance the user experience by taking user emotions into consideration.

[0324] Hardware and software used

[0325] This system mainly uses the following hardware and software:

[0326] Server: A high-performance server for data analysis and processing. Specific examples include server instances on AWS or Google Cloud Platform.

[0327] Terminal: A device operated by a user. Examples include PCs, smartphones, tablets, etc.

[0328] Emotion engine: Software for analyzing a user's facial expressions and voice data. Examples include OpenCV and Microsoft® Azure® Emotion API.

[0329] Generative AI model: An artificial intelligence model for content generation. An example of this is OpenAI GPT-3 (registered trademark).

[0330] Natural Language Processing (NLP) software, used to analyze generated content and check for inappropriate content and copyright infringement risks. Examples include spaCy and NLTK.

[0331] Specific operation of the system

[0332] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expressions and voice data through the connected camera and microphone. The collected data is analyzed in real time to recognize the user's emotions.

[0333] The user fills out a content generation request form for the AI ​​on their device screen and clicks the "Submit" button. The request includes the type of content they want to generate and guidelines.

[0334] The server receives content generation requests sent by users, formats them as necessary, and forwards them to the generative AI model, which generates content based on the requests and meets the specified conditions.

[0335] The server receives the content generated by the generative AI model, analyzes it using NLP software, and blocks it if it contains inappropriate content or is at risk of copyright infringement.

[0336] The emotion engine analyzes the user's emotional state and determines whether the content is appropriate for the user's state. If necessary, it adjusts the generated content. For example, if the user is feeling stressed, it changes the content to something more relaxing.

[0337] Finally, the server checks the generated content for appropriateness and provides it to the user. If the content is inappropriate, the server notifies the user of the result.

[0338] Specific examples

[0339] As an example, consider the case where a user submits a content generation request using the following prompt:

[0340] Example prompt sentence:

[0341] "Generate fun adventure stories for children, but please do not include dangerous or violent scenes."

[0342] Based on this request, a generative AI model generates an adventure story. The server analyzes the content to ensure it does not contain any dangerous scenes. Furthermore, an emotion engine analyzes the user's emotional state and adjusts the content as needed. The optimal content is then delivered to the user.

[0343] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0344] Step 1:

[0345] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expression and voice data through the connected camera and microphone and analyzes it in real time. The input data is the user's facial expression and voice data, and the output is the analysis result that indicates the user's emotional state. Specifically, it calls the emotion engine's API and analyzes the collected data.

[0346] Step 2:

[0347] The user fills out a content generation request form for the generation AI on their device screen and clicks the "Submit" button. The input data is form data that includes the type of content they want to generate and guidelines, and the output data is the request data that is sent to the server. Specifically, the user fills out the form on the browser screen and sends a POST request.

[0348] Step 3:

[0349] The server receives a content generation request sent by the user. The received data is the user's request data, and the output is formatted request data to be sent to the generative AI model. The server formats the received request content as needed and sends it to the generative AI model. Specifically, it converts the received data into an internal format and sends it to the generative AI's API.

[0350] Step 4:

[0351] The generation AI generates content that meets the specified conditions based on the request. The input is the request data transferred from the server, and the output is the generated content data. Specifically, the generation AI runs a natural language generation model (e.g., GPT-3) based on the prompt sentence and outputs the generated results as text data.

[0352] Step 5:

[0353] The server receives the content generated by the generation AI. The input is the content data from the generation AI, and the output is the data to be analyzed for filtering and moderation. Specifically, the server receives the HTTP response and stores the generated content data in an internal database.

[0354] Step 6:

[0355] The server automatically filters and moderates the content it receives. The input is the content data to be analyzed, and the output is the filtering and moderation results. Specifically, it uses an NLP library (e.g., spaCy or NLTK) to analyze the text and check for inappropriate content and copyright infringement risks.

[0356] Step 7:

[0357] Based on the results of the emotion engine, the server determines whether the content is appropriate for the user's emotional state and adjusts the content as necessary. The input is the emotion recognition result and the content data to be analyzed, and the output is the adjusted content data. Specifically, the emotion recognition result is obtained from the API, and the content is regenerated or replaced with a template as necessary.

[0358] Step 8:

[0359] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it blocks the content. The input is the filtering and moderation results, and the output is a list of blocked content. The specific operation is to add the content to the block list so that it is not provided to users.

[0360] Step 9:

[0361] The server notifies the user whether the generated content is appropriate or blocked because it is inappropriate. The input is the user's notification destination information and notification content data, and the output is the notification sending result. Specifically, the server notifies the user of the result via email or an in-app notification system.

[0362] (Application example 2)

[0363] 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."

[0364] Content generated by conventional generative AI models is not optimized based on user sentiment, and therefore does not provide a sufficiently good user experience. Furthermore, content may contain inappropriate content or risk copyright infringement, making it difficult to provide safe and appropriate content. This has led to problems such as reduced user satisfaction.

[0365] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0366] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking generated content if it contains inappropriate content, means for notifying users about the generated content, means for analyzing emotions of users, and means for adjusting the content of the generated content based on the emotions of users, thereby enabling the provision of appropriate content that takes into consideration the emotions of users.

[0367] "Means for automatically filtering generated content" refers to technology that automatically analyzes generated content and detects and removes inappropriate content or harmful information.

[0368] The "means for assessing the risk of copyright infringement of generated content" is a technology for assessing whether generated content uses existing copyrighted works without permission and determining the risk of copyright infringement.

[0369] "Means for blocking generated content if it contains inappropriate content" refers to technology that removes or stops the display of generated content if it contains inappropriate language or harmful information.

[0370] The "means for notifying users of generated content" is a technique for notifying users whether generated content is in an appropriate state when provided to the users.

[0371] "Emotion analysis means for recognizing user emotions" is a technology that analyzes the user's facial expressions, tone of voice, input data, etc. to recognize the user's emotional state in real time.

[0372] The "means for adjusting the content of generated content based on the user's emotions" is a technology for optimizing the content of generated content in accordance with the recognized user's emotions.

[0373] System Overview

[0374] The system for implementing this invention integrates multiple key functions, analyzes user emotions in real time, and appropriately filters and adjusts the generated content to provide high-quality content.

[0375] Program processing

[0376] The program of this system includes the following elements:

[0377] 1. Starting and analyzing the emotion recognition engine

[0378] The server collects data in real time from the camera and microphone of the user's device (smartphone, tablet, etc.), which allows it to analyze the user's facial expressions and voice and recognize their emotions.

[0379] The software used is a general image analysis API (e.g., image analysis cloud service) for image analysis, and an audio tone analysis tool (e.g., audio tone analysis service) for audio analysis.

[0380] 2. Submitting a content generation request

[0381] The user sends a request to the AI ​​via their device, including the type of content they want and keywords, and the request includes a specific prompt.

[0382] 3. Content generation using generative AI models

[0383] The server generates content based on specified conditions based on a generative AI model (e.g., a generative AI service).

[0384] The generated content is diverse, including text, images, and audio.

[0385] Specific examples

[0386] For example, if a user wants to send a request to create relaxing music, they might use a prompt like this:

[0387] Example prompt: "Generate relaxing music containing nature sounds that users would like to listen to while in a relaxed state."

[0388] Based on this request, the server issues instructions to the generation AI to generate the requested content.

[0389] 4. Filtering and Moderation

[0390] The server receives the generated content and automatically checks it for inappropriate content and risks of copyright infringement.

[0391] It uses natural language processing (NLP) to analyze text and uses machine learning libraries such as TENSORFLOW® and PyTorch.

[0392] 5. Emotion-based content adjustment

[0393] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, the server regenerates content to provide soothing sensations without overstimulating them.

[0394] The parameters of the AI ​​model are adjusted based on the results of sentiment analysis to generate new content.

[0395] 6. Provision of Final Content and Notification

[0396] The server provides the tailored content to the user and notifies them if the content is appropriate.

[0397] If the user is in a relaxed state, content with a high relaxation effect is provided, improving the user experience.

[0398] As described above, the present invention takes into consideration the feelings of users and makes it possible to provide optimal content while reducing the risk of inappropriate content and copyright infringement.

[0399] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0400] Step 1:

[0401] The server collects data in real time from the camera and microphone of the device (smartphone, tablet, etc.). This data is used to obtain the user's facial expressions and voice data, and preprocesses it to recognize emotions. The collected data is analyzed using an image analysis cloud service and a voice tone analysis service, and the user's emotional state (e.g., "relaxed") is output.

[0402] Step 2:

[0403] The user inputs and sends a request through their device, including the type of content they want and keywords. The generated prompt is something like "Generate relaxing music." The server receives this request and prepares the prompt as data to send appropriate instructions to the generative AI model.

[0404] Step 3:

[0405] The server sends a prompt to the generative AI model, instructing it to generate content based on the specified conditions. The generative AI model generates content such as text, images, and audio based on the request. This generated content is then sent back to the server.

[0406] Step 4:

[0407] The server receives the generated content and automatically filters it for inappropriate content and copyright infringement risks. It uses natural language processing (NLP) to analyze the text and machine learning libraries (e.g., TensorFlow and PyTorch) to assess the risk. Based on this assessment, it blocks the content if it contains inappropriate content.

[0408] Step 5:

[0409] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, it checks whether the generated content is overly stimulating and regenerates it if necessary. It also adjusts the parameters of the AI ​​model to optimize the newly generated content.

[0410] Step 6:

[0411] The server provides the final adjusted content to the user and notifies them whether the content is appropriate. As the content is provided, the emotion engine analyzes it again in light of the user's emotions and provides appropriate feedback. If the user is relaxed, relaxation content is provided to maintain that state.

[0412] As described above, this system specifically realizes a mechanism for analyzing user emotions in real time and generating and providing optimal content based on that.

[0413] 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.

[0414] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (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.

[0415] 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.

[0416] [Second embodiment]

[0417] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0418] 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.

[0419] 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).

[0420] 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.

[0421] 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.

[0422] 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).

[0423] 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.

[0424] 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.

[0425] 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.

[0426] 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.

[0427] In the smart glasses 214, 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.

[0428] 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."

[0429] The present invention relates to a system for automatically filtering and moderating content generated by generative AI to reduce the risk of inappropriate content and copyright infringement. The system provides a means for analyzing the generated content and detecting the risk of inappropriate content and copyright infringement.

[0430] System Overview

[0431] The system includes the following main functions:

[0432] 1. Automatically filtering generated content

[0433] 2. Means of assessing the risk of copyright infringement of generated content

[0434] 3. How to block generated content if it contains inappropriate content

[0435] 4. Means of notifying users about generated content

[0436] Program processing

[0437] The program of this system performs processing in the following procedure.

[0438] 1. Submitting a content generation request

[0439] The user uses a device to send a content generation request to the generation AI, which includes the type of content they want to generate and guidelines.

[0440] 2. Receiving and forwarding requests

[0441] The server receives the request from the user and forwards the request to the generation AI.

[0442] 3. Content generation using generative AI

[0443] The generation AI generates the specified content based on the request.

[0444] 4. Receiving Generated Content

[0445] The server receives the generated content from the generation AI.

[0446] 5. Start filtering and moderation

[0447] Server-generated content is automatically filtered and moderated, using natural language processing (NLP) to analyze text and detect inappropriate language, and it also compares content similarity with existing databases to assess the risk of copyright infringement.

[0448] 6. Blocking inappropriate content

[0449] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[0450] 7. Notice to Users

[0451] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0452] Specific examples

[0453] 1. User Request Submission

[0454] A user submits a request to "generate an adventure story for children."

[0455] 2. Receiving and forwarding requests

[0456] The server receives the request and forwards it to the generation AI.

[0457] 3. Story Generation Using Generative AI

[0458] The AI ​​generates an adventure story in which a prince and princess encounter mysterious creatures in the forest.

[0459] 4. Receiving Generated Content

[0460] The server receives the generated story.

[0461] 5. Filtering and moderation processes

[0462] The server analyzes the story text and checks for inappropriate language and similarities to existing works. For example, if a story contains a scene in which the prince indiscriminately defeats monsters, it will automatically detect it as inappropriate.

[0463] 6. Blocking inappropriate content

[0464] The server detects inappropriate content and blocks the story.

[0465] 7. Notifying Users and Providing Results

[0466] The server sends a notification to the user saying "Generation blocked due to inappropriate content."

[0467] As a result, the present invention provides an environment in which users can use generative AI with peace of mind, improving the quality of generated content and reducing legal risks.

[0468] The processing flow will be explained below.

[0469] Program processing flow

[0470] Step 1:

[0471] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[0472] Step 2:

[0473] The server receives the user's request, which includes details about the content, and prepares to parse it and forward it to the generation AI in the appropriate format.

[0474] Step 3:

[0475] The server forwards the content generation request to the generation AI, which starts the process of the generation AI generating content based on the specified conditions.

[0476] Step 4:

[0477] Generative AI generates content based on user requests, such as "A story about a prince and princess having an adventure in the forest."

[0478] Step 5:

[0479] The server receives the generated content from the generation AI. This content is raw, as it has not yet been filtered or moderated.

[0480] Step 6:

[0481] The server then begins the filtering and moderation process on the generated content, first using natural language processing (NLP) to analyze the text and check for inappropriate language or expressions.

[0482] Step 7:

[0483] The server then compares the content similarity with existing databases to assess the risk of copyright infringement of the generated content. If similar content is found, the content is deemed to be at risk of copyright infringement.

[0484] Step 8:

[0485] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters would be deemed inappropriate.

[0486] Step 9:

[0487] The server generates a result indicating whether the content is appropriate or not and notifies the user. If the content is appropriate, it is served to the user as is, otherwise an error message is sent.

[0488] Step 10:

[0489] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they use it, and if it is inappropriate, they send a request again.

[0490] In this way, the system automatically checks the quality and legal risks of generated content, providing a safe and secure environment for users.

[0491] Example 1

[0492] 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."

[0493] Since content generated using generative AI may contain inappropriate content or may infringe copyrights, there is no environment in place for users to use it safely. For this reason, there is a need for a system that can automatically filter generated content and provide only appropriate content.

[0494] 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.

[0495] In this invention, the server includes: a means for a user to send a content generation request to the generation AI using a terminal; a means for the server to receive the request from the user and forward it to the generation AI; a means for the generation AI to generate content based on the request; a means for the server to receive the generated content from the generation AI; a means for the server to automatically filter and moderate the generated content; a means for the server to block the generated content if the server determines that the generated content contains inappropriate content or poses a risk of copyright infringement; and a means for the server to notify the user of the result of whether the generated content is appropriate or has been blocked. This makes it possible to improve the quality of generated content and reduce legal risks.

[0496] "User" means an individual or entity that uses a device to send a content generation request to a generation AI.

[0497] A "terminal" is an electronic device used by a user to send a content generation request to the generation AI, including a computer, smartphone, tablet, etc.

[0498] "Generative AI" is an artificial intelligence system that automatically generates content based on a given request.

[0499] A "server" is a computer system that receives requests from users via the Internet, forwards them to the generation AI, and receives and processes the generated content.

[0500] A "content generation request" is an instruction sent by a user to a generation AI to generate specific content, including the content and guidelines.

[0501] "Filtering" is the process of automatically detecting and filtering out inappropriate content from generated content.

[0502] "Moderation" is the process of evaluating and managing generated content to detect inappropriate content or the risk of copyright infringement.

[0503] "Notification" is the act of the server informing the user whether the generated content is appropriate or has been blocked because it is inappropriate.

[0504] "Natural Language Processing (NLP)" refers to the technology of analyzing, understanding, and appropriately processing human language. It is used for text analysis of generated content.

[0505] "Copyright infringement" is the act of using someone else's copyright without permission, which carries legal risks.

[0506] "Blocking" is the act of restricting access to prevent users from being provided with inappropriate content or generated content that poses a risk of copyright infringement.

[0507] This invention relates to a system that automatically filters and moderates content generated by generative AI, aiming to improve the quality of generated content and reduce legal risks. This system is composed of elements such as users, terminals, servers, and generative AI, and by clarifying the roles of each, it achieves efficient and safe content generation.

[0508] System Configuration

[0509] User

[0510] Users send content generation requests to the generative AI using a dedicated device, which can be a computer, smartphone, tablet, or other device, and access the generative AI model using a web browser or dedicated application.

[0511] Terminal

[0512] The terminal is a device through which users can send content generation requests to the generation AI, and acts as a user interface. A request form is displayed here, and users can enter the type of content they want to generate and guidelines.

[0513] server

[0514] The server receives user requests and forwards them appropriately to the AI ​​generator. It also receives the generated content returned by the AI ​​generator and automatically filters and moderates it. Specific software used in this process includes the natural language processing (NLP) libraries SpaCy and NLTK.

[0515] Generation AI

[0516] Generative AI is an artificial intelligence system that generates text content based on user requests, including language models like GPT-4.

[0517] Specific examples

[0518] 1. User Request Submission

[0519] A user sends a request from their device to "generate an adventure story for children." Specifically, they enter "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or pose a risk of copyright infringement" into the web form and click the submit button.

[0520] 2. Receiving and forwarding requests

[0521] The server receives the HTTP request, analyzes its contents, and converts it into JSON format for transmission to the API of the AI ​​generation. At this time, the request contents are transmitted to the endpoint of the AI ​​generation.

[0522] 3. Story Generation Using Generative AI

[0523] Based on the request, the AI ​​generates an "adventure story about a prince and princess meeting mysterious creatures in the forest." The generated content is sent back to the server in JSON format.

[0524] 4. Receiving and Processing Generated Content

[0525] The server receives the generated stories and stores them in an internal data structure, then uses an NLP library (e.g., SpaCy or NLTK) to analyze the text and check for profanity and similarity to existing works.

[0526] 5. Blocking inappropriate content

[0527] If the server determines that the content contains inappropriate content based on the analysis results, it will block the content. For example, if the content contains a scene in which the prince indiscriminately kills monsters, the filtering function will automatically detect it as inappropriate.

[0528] 6. Notice to Users

[0529] The server notifies the user whether the generated content is appropriate or has been blocked because it is inappropriate. Specifically, the user is shown a message saying, "The generation has been blocked because it contains inappropriate content."

[0530] Prompt Sentence Examples

[0531] Use the following prompt to input the generative AI model:

[0532] Create an adventure story for children. Please be careful not to include violent scenes or copyright infringement in the content.

[0533] This system allows users to use generative AI with confidence, reducing legal risks while producing high-quality content.

[0534] keyword

[0535] Generative AI model, prompt sentence

[0536] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0537] Step 1:

[0538] The user uses their device to send a content generation request to the generation AI. This request includes the type of content they want to generate and guidelines. Specifically, the user enters the following into a web form: "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or risk copyright infringement." and clicks the submit button. This form input is sent to the server as a request.

[0539] input:

[0540] User input into a web form (content generation request)

[0541] output:

[0542] HTTP request to the server

[0543] Step 2:

[0544] The server receives the request sent by the user. It analyzes the received request and converts it into a format suitable for the API of the generation AI (for example, JSON format). It then forwards the request to the endpoint of the generation AI. Specifically, the server checks the contents of the request, converts the format if necessary, and sends it to the API of the generation AI.

[0545] input:

[0546] HTTP request from the user

[0547] output:

[0548] The converted request to the generating AI (JSON format)

[0549] Step 3:

[0550] The generation AI generates content based on the received request. The generation AI uses an internal algorithm (e.g., GPT-4) to automatically generate text based on the input prompt. The generated content is sent back to the server in JSON format.

[0551] input:

[0552] The transformed request to the generative AI

[0553] output:

[0554] Generated content (JSON format)

[0555] Step 4:

[0556] The server receives the generated content from the generation AI, stores it in an internal data structure, and prepares it for the next filtering step. Specifically, the server parses the JSON-formatted data and extracts the text portion of the content.

[0557] input:

[0558] Generated content (JSON format)

[0559] output:

[0560] Storing text in internal data structures

[0561] Step 5:

[0562] The server automatically filters and moderates the generated content. During this process, it uses natural language processing (NLP) libraries (such as SpaCy or NLTK) to analyze the text and check for inappropriate content and similarities. Specifically, text analysis detects violent scenes and inappropriate content, and compares the generated content's similarities with existing databases. For example, if a video contains a scene in which a prince indiscriminately defeats monsters, it will be detected as violent.

[0563] input:

[0564] Generated content text

[0565] output:

[0566] Filtering and Moderation Results

[0567] Step 6:

[0568] If the server determines that the content is inappropriate or poses a risk of copyright infringement as a result of filtering and moderation, it will block the content. Specifically, if the filtering result is negative, the generated content will not be returned to the user and an error message will be prepared.

[0569] input:

[0570] Filtering and Moderation Results

[0571] output:

[0572] Generate appropriate content or error messages

[0573] Step 7:

[0574] The server notifies the user whether the generated content is appropriate or inappropriate and has been blocked. The server returns the generated content or an error message to the user's device as an HTTP response. Specifically, a message stating "Generation has been blocked because it contains inappropriate content" is displayed.

[0575] input:

[0576] Appropriate content or error messages

[0577] output:

[0578] HTTP response to the user

[0579] ---

[0580] The above is the specific flow of the program processing of this system. This allows users to use the generative AI with peace of mind, improves the quality of generated content, and reduces legal risks.

[0581] (Application example 1)

[0582] 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."

[0583] If the content generated by generative AI contains inappropriate content or poses a risk of copyright infringement, it is difficult to provide a safe environment for users. Furthermore, there is a lack of effective means to check for these risks before the generated content is distributed, raising concerns about legal risks and a decline in quality. In particular, in the case of educational content, the inclusion of inappropriate content not only reduces the educational effectiveness but also reduces the credibility of the content.

[0584] 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.

[0585] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying users of the generated content, means for checking the generated content before distribution, means for analyzing the audio and subtitles of the generated content using natural language processing, and means for evaluating the generated content based on guidelines specialized for educational content. This allows the generated content to be used and distributed safely without containing inappropriate content or the risk of copyright infringement.

[0586] "Generated content" refers to digital content generated by a generative AI in response to a user request.

[0587] "Filtering" refers to the process of automatically detecting and removing inappropriate language or content contained in generated content.

[0588] "Risk of copyright infringement" refers to situations in which generated content may infringe the rights of existing copyrighted works.

[0589] "Inappropriate content" refers to any expression or information that may violate user guidelines or the law.

[0590] "Notification to users" refers to a means of notifying users of the evaluation results and block information of generated content.

[0591] The "means for checking before distribution" refers to a process for confirming the appropriateness of generated content before distributing the content to a user or a third party.

[0592] "Natural language processing" refers to the technology that enables computers to understand, generate, and interpret human language.

[0593] "Means for analyzing audio and subtitles" refers to technology for automatically analyzing the audio data and subtitle data contained in the generated content and evaluating its content.

[0594] "Educational Content Specific Guidelines" means specific standards or rules that apply to generated content used for educational purposes.

[0595] This invention is a system for automatically filtering and moderating content generated by generative AI, with the aim of reducing the risk of inappropriate content and copyright infringement, particularly in the generation of educational content, by eliminating inappropriate content and providing reliable, high-quality content.

[0596] System Configuration

[0597] The system consists of the following main components:

[0598] 1. Server: Receives requests for generated content and forwards them to the generation AI. Receives the generated content and performs filtering and moderation. Hardware used could be an AWS EC2 instance, for example.

[0599] 2. Generative AI: Generates content based on user requests, for example using OpenAI's GPT-4.

[0600] 3. Natural language processing tools: Analyze the audio and subtitles of generated content to assess the risk of inappropriate content and copyright infringement. For example, Google Cloud Natural Language API is used.

[0601] 4. User device: Sends requests and receives generated content. Uses a smartphone (iOS or Android device).

[0602] Data processing and calculation

[0603] Receiving and forwarding requests: The server forwards requests for generated content received from the user device to the generation AI. The request includes the type of content to be generated and guidelines.

[0604] Content Generation: Generative AI generates content based on given guidelines, for example, given a prompt to generate an "educational video learning about the minerals of the Earth."

[0605] Filtering and moderation: The server automatically filters and moderates generated content, using natural language processing (Google Cloud Natural Language API) to analyze text and audio data and assess risk of inappropriate content and copyright infringement.

[0606] Notification to users: Based on the evaluation results, users will be notified whether the generated content is appropriate. If it is deemed inappropriate, users will be notified along with the reason.

[0607] Specific examples

[0608] A user might submit a request to generate educational video content, for example, "Please generate an educational video about the minerals of the Earth. Please do not include any inappropriate language or copyright infringing content."

[0609] The server forwards this request to the generation AI, which generates an educational video according to the specified guidelines. The server then receives the generated video and analyzes the video's audio and subtitles using natural language processing tools. If the analysis finds that the content does not comply with the educational guidelines or poses a risk of copyright infringement, the video is blocked. Finally, the server notifies the user of the appropriateness of the generated content.

[0610] In this way, users can confidently create and distribute high-quality, reliable educational content.

[0611] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0612] Step 1:

[0613] The user uses their device to send a request to the AI. The request includes the type of content they want to generate and guidelines. For example, they might enter a prompt like, "Generate an educational video that teaches about the minerals on Earth. Please do not include any inappropriate language or copyright infringement." This request is then sent to the server.

[0614] Step 2:

[0615] The server parses the request received from the user and converts it into an appropriate format. The server then forwards the request to the generation AI, which sends the request data to a specific API endpoint and applies prompts so that the generation AI can understand the content.

[0616] Step 3:

[0617] The generation AI generates content based on the request. The generation AI receives prompt text as input and outputs generated content such as an educational video script or narration. Each generation AI has different parameters and settings, so processing is done in the appropriate data format.

[0618] Step 4:

[0619] The server receives the generated content from the generation AI, records its contents, verifies the data format of the generated content (video script and narration), and converts it to fit the next processing step.

[0620] Step 5:

[0621] The server analyzes the generated content using natural language processing (NLP) tools. Audio and subtitle data is converted into text format and analyzed through a natural language processing engine (e.g., Google Cloud Natural Language API). The results of this analysis are used to assess the risk of inappropriate content and copyright infringement.

[0622] Step 6:

[0623] The server filters the generated content based on the analysis results. If inappropriate content or a risk of copyright infringement is detected, the server blocks the content. If necessary, the server modifies part of the generated content or requests the AI ​​to regenerate it.

[0624] Step 7:

[0625] The server notifies the user of the evaluation results of the generated content. If the content is appropriate, the server notifies the user and provides a download link and notification that the content is ready for distribution. If the content contains inappropriate content, the server notifies the user, along with the reason for the inappropriate content.

[0626] In this way, through a series of processing steps, a system is realized that ensures the quality of user-generated content and provides an environment in which it can be distributed safely.

[0627] 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.

[0628] This invention relates to a system that provides more appropriate content by combining a system that reduces the risk of inappropriate content and copyright infringement in content generated by generative AI with an emotion engine that recognizes user emotions. By taking user emotions into consideration, this system aims to further improve the quality of generated content and enhance the user experience.

[0629] System Overview

[0630] The system includes the following main functions:

[0631] 1. Automatically filtering generated content

[0632] 2. Means of assessing the risk of copyright infringement of generated content

[0633] 3. How to block generated content if it contains inappropriate content

[0634] 4. Means of notifying users about generated content

[0635] 5. Emotion engine that recognizes user emotions

[0636] 6. The ability to adjust generated content based on user sentiment

[0637] 7. Means for suspending provision of generated content based on emotion recognition results

[0638] Program processing

[0639] The program of this system performs processing in the following procedure.

[0640] 1. Activating the Emotional Engine

[0641] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time. This analysis recognizes the user's emotions.

[0642] 2. Submitting a content generation request

[0643] The user sends a content generation request to the AI ​​via their device, which includes the type of content they want to generate and guidelines.

[0644] 3. Receiving and forwarding requests

[0645] The server receives the user's request and forwards it to the generation AI.

[0646] 4. Content generation using generative AI

[0647] The generation AI generates content that meets the specified conditions based on the request.

[0648] 5. Receiving Generated Content

[0649] The server receives the generated content from the generation AI.

[0650] 6. Start filtering and moderation

[0651] Automatic filtering and moderation of server-generated content, using natural language processing (NLP) to analyze text and check for inappropriate language and copyright infringement.

[0652] 7. Emotion-aware content adjustment

[0653] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state and adjusts the content as necessary. For example, if the user is feeling stressed, the server may provide a relaxing story.

[0654] 8. Blocking inappropriate content

[0655] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[0656] 9. Notice to Users

[0657] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0658] Specific examples

[0659] 1. User Emotion Recognition

[0660] While the user is viewing content on their mobile device, the emotion engine analyzes the user's facial expression data in real time and recognizes their "enjoyment" state.

[0661] 2. Sending a content generation request

[0662] A user submits a request to "generate an adventure story for children."

[0663] 3. Adventure Story Generation Using Generative AI

[0664] The AI ​​generates an "adventure story," which includes dangerous adventure scenes.

[0665] 4. Filtering and Moderation

[0666] The server performs filtering and moderation, determining that dangerous scenes are inappropriate and blocking them.

[0667] 5. Adjustment based on emotion recognition results

[0668] In order to keep the user in a "fun" state, the emotion engine adjusts to prioritize the generation of new stories based on "fun adventures."

[0669] 6. Delivery of Final Content

[0670] The server generates a tailored and appropriate adventure story and provides it to the user.

[0671] In this way, the system can provide appropriate content while taking into account the user's emotions, improving the overall user experience.

[0672] The processing flow will be explained below.

[0673] Specific processing flow of the program

[0674] Step 1:

[0675] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time, thereby recognizing the user's emotional state.

[0676] Step 2:

[0677] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[0678] Step 3:

[0679] The server receives the user's request, analyzes the request content, and prepares to forward it to the generation AI in an appropriate format.

[0680] Step 4:

[0681] The server forwards a content generation request to the generation AI, and the generation AI starts generating content based on the specified conditions.

[0682] Step 5:

[0683] The generation AI generates the specified content based on the request, for example, "A story about a prince and princess having an adventure in the forest."

[0684] Step 6:

[0685] The server receives the generated content from the generative AI, which has not yet been filtered or moderated.

[0686] Step 7:

[0687] The server analyzes the generated content using natural language processing (NLP) to check for inappropriate words and expressions.

[0688] Step 8:

[0689] The server evaluates the risk of copyright infringement of generated content by matching the content similarity with existing databases. If similar content is found, the content is determined to be at risk of copyright infringement.

[0690] Step 9:

[0691] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state. For example, if the user is feeling anxious, the content can be adjusted to change it into a relaxing story.

[0692] Step 10:

[0693] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters is deemed inappropriate and blocked.

[0694] Step 11:

[0695] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0696] Step 12:

[0697] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they can use it, and if it is inappropriate, they are given the option to send the request again or receive the adjusted content.

[0698] Specific examples

[0699] 1. Step 1-6: The user sends a request to "generate an adventure story for children," and the server receives the request and forwards it to the generation AI. The generation AI generates a story about a prince and princess having an adventure in the forest, and the server receives it.

[0700] 2. Step 7: The server analyzes the generated story using natural language processing, for example, to check whether it contains any "dangerous scenes."

[0701] 3. Step 8: The server checks the generated content against an existing database to assess the risk of copyright infringement.

[0702] 4. Step 9: The emotion engine recognizes the user's "anxiety" state, and the server adjusts the story content to a "relaxing adventure story" based on this result.

[0703] 5. Steps 10-12: The server determines that the final adjusted story is appropriate and notifies the user, who then accepts and views the story.

[0704] In this way, the system automatically checks the quality and legal risks of generated content, and further adjusts the content based on user sentiment, providing an environment where users can use it with peace of mind.

[0705] Example 2

[0706] 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."

[0707] The content generated by modern generative AI models may contain inappropriate content or risk copyright infringement. Furthermore, if the generated content does not match the user's emotional state, it can negatively impact the user experience. To address this issue, it is necessary to ensure the appropriateness of the content and take the user's emotional state into consideration.

[0708] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying the user about the generated content, means for recognizing the user's emotion, means for adjusting the content of the generated content based on the user's emotion, and means for suspending provision of the generated content based on the result of the user's emotion recognition. This makes it possible to provide content that matches the user's emotional state while ensuring the appropriateness of the generated content.

[0709] "Generated content" refers to information such as text, images, and video that is automatically generated by a generative AI model based on a user request.

[0710] "Filtering measures" are technologies or algorithms used to analyze generated content and determine whether it contains inappropriate content or is at risk of copyright infringement.

[0711] "Copyright infringement risk assessment means" means any technology or algorithm that compares generated content with existing databases or copyrighted material to assess the risk of copyright infringement.

[0712] "Blocking measures" are technologies or methods that remove generated content before it is provided to users if the content contains inappropriate content or poses a risk of copyright infringement.

[0713] A "notification mechanism" is a mechanism for informing users of the outcome of generated content, whether it is appropriate or blocked for being inappropriate.

[0714] "Emotion recognition means" refers to technology or algorithms that analyze a user's facial expressions, voice, and input data, and recognize the user's emotions in real time based on the results.

[0715] The "emotion-based content adjustment means" refers to a technology or algorithm for changing the content of generated content to match the emotional state of the user based on the result of the user's emotion recognition.

[0716] The "provision interruption means" is a technique or method for interrupting the provision of generated content midway when the user's emotion recognition result satisfies certain conditions.

[0717] MODE FOR CARRYING OUT THE INVENTION

[0718] This invention is a system that reduces the risk of inappropriate content and copyright infringement in content generated using a generative AI model, and combines it with an emotion engine that recognizes user emotions to provide more appropriate content. This system aims to further improve the quality of generated content and enhance the user experience by taking user emotions into consideration.

[0719] Hardware and software used

[0720] This system mainly uses the following hardware and software:

[0721] Server: A high-performance server for data analysis and processing. Specific examples include server instances on AWS or Google Cloud Platform.

[0722] Terminal: A device operated by a user. Examples include PCs, smartphones, tablets, etc.

[0723] Emotion engine: Software for analyzing a user's facial expressions and voice data. Examples include OpenCV and Microsoft Azure Emotion API.

[0724] Generative AI model: An artificial intelligence model for content generation. An example of this is OpenAI GPT-3.

[0725] Natural Language Processing (NLP) software, used to analyze generated content and check for inappropriate content and copyright infringement risks. Examples include spaCy and NLTK.

[0726] Specific operation of the system

[0727] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expressions and voice data through the connected camera and microphone. The collected data is analyzed in real time to recognize the user's emotions.

[0728] The user fills out a content generation request form for the AI ​​on their device screen and clicks the "Submit" button. The request includes the type of content they want to generate and guidelines.

[0729] The server receives content generation requests sent by users, formats them as necessary, and forwards them to the generative AI model, which generates content based on the requests and meets the specified conditions.

[0730] The server receives the content generated by the generative AI model, analyzes it using NLP software, and blocks it if it contains inappropriate content or is at risk of copyright infringement.

[0731] The emotion engine analyzes the user's emotional state and determines whether the content is appropriate for the user's state. If necessary, it adjusts the generated content. For example, if the user is feeling stressed, it changes the content to something more relaxing.

[0732] Finally, the server checks the generated content for appropriateness and provides it to the user. If the content is inappropriate, the server notifies the user of the result.

[0733] Specific examples

[0734] As an example, consider the case where a user submits a content generation request using the following prompt:

[0735] Example prompt sentence:

[0736] "Generate fun adventure stories for children, but please do not include dangerous or violent scenes."

[0737] Based on this request, a generative AI model generates an adventure story. The server analyzes the content to ensure it does not contain any dangerous scenes. Furthermore, an emotion engine analyzes the user's emotional state and adjusts the content as needed. The optimal content is then delivered to the user.

[0738] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0739] Step 1:

[0740] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expression and voice data through the connected camera and microphone and analyzes it in real time. The input data is the user's facial expression and voice data, and the output is the analysis result that indicates the user's emotional state. Specifically, it calls the emotion engine's API and analyzes the collected data.

[0741] Step 2:

[0742] The user fills out a content generation request form for the generation AI on their device screen and clicks the "Submit" button. The input data is form data that includes the type of content they want to generate and guidelines, and the output data is the request data that is sent to the server. Specifically, the user fills out the form on the browser screen and sends a POST request.

[0743] Step 3:

[0744] The server receives a content generation request sent by the user. The received data is the user's request data, and the output is formatted request data to be sent to the generative AI model. The server formats the received request content as needed and sends it to the generative AI model. Specifically, it converts the received data into an internal format and sends it to the generative AI's API.

[0745] Step 4:

[0746] The generation AI generates content that meets the specified conditions based on the request. The input is the request data transferred from the server, and the output is the generated content data. Specifically, the generation AI runs a natural language generation model (e.g., GPT-3) based on the prompt sentence and outputs the generated results as text data.

[0747] Step 5:

[0748] The server receives the content generated by the generation AI. The input is the content data from the generation AI, and the output is the data to be analyzed for filtering and moderation. Specifically, the server receives the HTTP response and stores the generated content data in an internal database.

[0749] Step 6:

[0750] The server automatically filters and moderates the content it receives. The input is the content data to be analyzed, and the output is the filtering and moderation results. Specifically, it uses an NLP library (e.g., spaCy or NLTK) to analyze the text and check for inappropriate content and copyright infringement risks.

[0751] Step 7:

[0752] Based on the results of the emotion engine, the server determines whether the content is appropriate for the user's emotional state and adjusts the content as necessary. The input is the emotion recognition result and the content data to be analyzed, and the output is the adjusted content data. Specifically, the emotion recognition result is obtained from the API, and the content is regenerated or replaced with a template as necessary.

[0753] Step 8:

[0754] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it blocks the content. The input is the filtering and moderation results, and the output is a list of blocked content. The specific operation is to add the content to the block list so that it is not provided to users.

[0755] Step 9:

[0756] The server notifies the user whether the generated content is appropriate or blocked because it is inappropriate. The input is the user's notification destination information and notification content data, and the output is the notification sending result. Specifically, the server notifies the user of the result via email or an in-app notification system.

[0757] (Application example 2)

[0758] 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."

[0759] Content generated by conventional generative AI models is not optimized based on user sentiment, and therefore does not provide a sufficiently good user experience. Furthermore, content may contain inappropriate content or risk copyright infringement, making it difficult to provide safe and appropriate content. This has led to problems such as reduced user satisfaction.

[0760] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0761] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking generated content if it contains inappropriate content, means for notifying users about the generated content, means for analyzing emotions of users, and means for adjusting the content of the generated content based on the emotions of users, thereby enabling the provision of appropriate content that takes into consideration the emotions of users.

[0762] "Means for automatically filtering generated content" refers to technology that automatically analyzes generated content and detects and removes inappropriate content or harmful information.

[0763] The "means for assessing the risk of copyright infringement of generated content" is a technology for assessing whether generated content uses existing copyrighted works without permission and determining the risk of copyright infringement.

[0764] "Means for blocking generated content if it contains inappropriate content" refers to technology that removes or stops the display of generated content if it contains inappropriate language or harmful information.

[0765] The "means for notifying users of generated content" is a technique for notifying users whether generated content is in an appropriate state when provided to the users.

[0766] "Emotion analysis means for recognizing user emotions" is a technology that analyzes the user's facial expressions, tone of voice, input data, etc. to recognize the user's emotional state in real time.

[0767] The "means for adjusting the content of generated content based on the user's emotions" is a technology for optimizing the content of generated content in accordance with the recognized user's emotions.

[0768] System Overview

[0769] The system for implementing this invention integrates multiple key functions, analyzes user emotions in real time, and appropriately filters and adjusts the generated content to provide high-quality content.

[0770] Program processing

[0771] The program of this system includes the following elements:

[0772] 1. Starting and analyzing the emotion recognition engine

[0773] The server collects data in real time from the camera and microphone of the user's device (smartphone, tablet, etc.), which allows it to analyze the user's facial expressions and voice and recognize their emotions.

[0774] The software used is a general image analysis API (e.g., image analysis cloud service) for image analysis, and an audio tone analysis tool (e.g., audio tone analysis service) for audio analysis.

[0775] 2. Submitting a content generation request

[0776] The user sends a request to the AI ​​via their device, including the type of content they want and keywords, and the request includes a specific prompt.

[0777] 3. Content generation using generative AI models

[0778] The server generates content based on specified conditions based on a generative AI model (e.g., a generative AI service).

[0779] The generated content is diverse, including text, images, and audio.

[0780] Specific examples

[0781] For example, if a user wants to send a request to create relaxing music, they might use a prompt like this:

[0782] Example prompt: "Generate relaxing music containing nature sounds that users would like to listen to while in a relaxed state."

[0783] Based on this request, the server issues instructions to the generation AI to generate the requested content.

[0784] 4. Filtering and Moderation

[0785] The server receives the generated content and automatically checks it for inappropriate content and risks of copyright infringement.

[0786] It utilizes natural language processing (NLP) to analyze text and uses machine learning libraries such as TensorFlow and PyTorch.

[0787] 5. Emotion-based content adjustment

[0788] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, the server regenerates content to provide soothing sensations without overstimulating them.

[0789] The parameters of the AI ​​model are adjusted based on the results of sentiment analysis to generate new content.

[0790] 6. Provision of Final Content and Notification

[0791] The server provides the tailored content to the user and notifies them if the content is appropriate.

[0792] If the user is in a relaxed state, content with a high relaxation effect is provided, improving the user experience.

[0793] As described above, the present invention takes into consideration the feelings of users and makes it possible to provide optimal content while reducing the risk of inappropriate content and copyright infringement.

[0794] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0795] Step 1:

[0796] The server collects data in real time from the camera and microphone of the device (smartphone, tablet, etc.). This data is used to obtain the user's facial expressions and voice data, and preprocesses it to recognize emotions. The collected data is analyzed using an image analysis cloud service and a voice tone analysis service, and the user's emotional state (e.g., "relaxed") is output.

[0797] Step 2:

[0798] The user inputs and sends a request through their device, including the type of content they want and keywords. The generated prompt is something like "Generate relaxing music." The server receives this request and prepares the prompt as data to send appropriate instructions to the generative AI model.

[0799] Step 3:

[0800] The server sends a prompt to the generative AI model, instructing it to generate content based on the specified conditions. The generative AI model generates content such as text, images, and audio based on the request. This generated content is then sent back to the server.

[0801] Step 4:

[0802] The server receives the generated content and automatically filters it for inappropriate content and copyright infringement risks. It uses natural language processing (NLP) to analyze the text and machine learning libraries (e.g., TensorFlow and PyTorch) to assess the risk. Based on this assessment, it blocks the content if it contains inappropriate content.

[0803] Step 5:

[0804] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, it checks whether the generated content is overly stimulating and regenerates it if necessary. It also adjusts the parameters of the AI ​​model to optimize the newly generated content.

[0805] Step 6:

[0806] The server provides the final adjusted content to the user and notifies them whether the content is appropriate. As the content is provided, the emotion engine analyzes it again in light of the user's emotions and provides appropriate feedback. If the user is relaxed, relaxation content is provided to maintain that state.

[0807] As described above, this system specifically realizes a mechanism for analyzing user emotions in real time and generating and providing optimal content based on that.

[0808] 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.

[0809] 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.

[0810] 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.

[0811] [Third embodiment]

[0812] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0813] 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.

[0814] 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).

[0815] 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.

[0816] 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.

[0817] 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).

[0818] 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.

[0819] 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.

[0820] 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.

[0821] 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.

[0822] 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.

[0823] 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."

[0824] The present invention relates to a system for automatically filtering and moderating content generated by generative AI to reduce the risk of inappropriate content and copyright infringement. The system provides a means for analyzing the generated content and detecting the risk of inappropriate content and copyright infringement.

[0825] System Overview

[0826] The system includes the following main functions:

[0827] 1. Automatically filtering generated content

[0828] 2. Means of assessing the risk of copyright infringement of generated content

[0829] 3. How to block generated content if it contains inappropriate content

[0830] 4. Means of notifying users about generated content

[0831] Program processing

[0832] The program of this system performs processing in the following procedure.

[0833] 1. Submitting a content generation request

[0834] The user uses a device to send a content generation request to the generation AI, which includes the type of content they want to generate and guidelines.

[0835] 2. Receiving and forwarding requests

[0836] The server receives the request from the user and forwards the request to the generation AI.

[0837] 3. Content generation using generative AI

[0838] The generation AI generates the specified content based on the request.

[0839] 4. Receiving Generated Content

[0840] The server receives the generated content from the generation AI.

[0841] 5. Start filtering and moderation

[0842] Server-generated content is automatically filtered and moderated, using natural language processing (NLP) to analyze text and detect inappropriate language, and it also compares content similarity with existing databases to assess the risk of copyright infringement.

[0843] 6. Blocking inappropriate content

[0844] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[0845] 7. Notice to Users

[0846] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[0847] Specific examples

[0848] 1. User Request Submission

[0849] A user submits a request to "generate an adventure story for children."

[0850] 2. Receiving and forwarding requests

[0851] The server receives the request and forwards it to the generation AI.

[0852] 3. Story Generation Using Generative AI

[0853] The AI ​​generates an adventure story in which a prince and princess encounter mysterious creatures in the forest.

[0854] 4. Receiving Generated Content

[0855] The server receives the generated story.

[0856] 5. Filtering and moderation processes

[0857] The server analyzes the story text and checks for inappropriate language and similarities to existing works. For example, if a story contains a scene in which the prince indiscriminately defeats monsters, it will automatically detect it as inappropriate.

[0858] 6. Blocking inappropriate content

[0859] The server detects inappropriate content and blocks the story.

[0860] 7. Notifying Users and Providing Results

[0861] The server sends a notification to the user saying "Generation blocked due to inappropriate content."

[0862] As a result, the present invention provides an environment in which users can use generative AI with peace of mind, improving the quality of generated content and reducing legal risks.

[0863] The processing flow will be explained below.

[0864] Program processing flow

[0865] Step 1:

[0866] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[0867] Step 2:

[0868] The server receives the user's request, which includes details about the content, and prepares to parse it and forward it to the generation AI in the appropriate format.

[0869] Step 3:

[0870] The server forwards the content generation request to the generation AI, which starts the process of the generation AI generating content based on the specified conditions.

[0871] Step 4:

[0872] Generative AI generates content based on user requests, such as "A story about a prince and princess having an adventure in the forest."

[0873] Step 5:

[0874] The server receives the generated content from the generation AI. This content is raw, as it has not yet been filtered or moderated.

[0875] Step 6:

[0876] The server then begins the filtering and moderation process on the generated content, first using natural language processing (NLP) to analyze the text and check for inappropriate language or expressions.

[0877] Step 7:

[0878] The server then compares the content similarity with existing databases to assess the risk of copyright infringement of the generated content. If similar content is found, the content is deemed to be at risk of copyright infringement.

[0879] Step 8:

[0880] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters would be deemed inappropriate.

[0881] Step 9:

[0882] The server generates a result indicating whether the content is appropriate or not and notifies the user. If the content is appropriate, it is served to the user as is, otherwise an error message is sent.

[0883] Step 10:

[0884] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they use it, and if it is inappropriate, they send a request again.

[0885] In this way, the system automatically checks the quality and legal risks of generated content, providing a safe and secure environment for users.

[0886] Example 1

[0887] 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."

[0888] Since content generated using generative AI may contain inappropriate content or may infringe copyrights, there is no environment in place for users to use it safely. For this reason, there is a need for a system that can automatically filter generated content and provide only appropriate content.

[0889] 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.

[0890] In this invention, the server includes: a means for a user to send a content generation request to the generation AI using a terminal; a means for the server to receive the request from the user and forward it to the generation AI; a means for the generation AI to generate content based on the request; a means for the server to receive the generated content from the generation AI; a means for the server to automatically filter and moderate the generated content; a means for the server to block the generated content if the server determines that the generated content contains inappropriate content or poses a risk of copyright infringement; and a means for the server to notify the user of the result of whether the generated content is appropriate or has been blocked. This makes it possible to improve the quality of generated content and reduce legal risks.

[0891] "User" means an individual or entity that uses a device to send a content generation request to a generation AI.

[0892] A "terminal" is an electronic device used by a user to send a content generation request to the generation AI, including a computer, smartphone, tablet, etc.

[0893] "Generative AI" is an artificial intelligence system that automatically generates content based on a given request.

[0894] A "server" is a computer system that receives requests from users via the Internet, forwards them to the generation AI, and receives and processes the generated content.

[0895] A "content generation request" is an instruction sent by a user to a generation AI to generate specific content, including the content and guidelines.

[0896] "Filtering" is the process of automatically detecting and filtering out inappropriate content from generated content.

[0897] "Moderation" is the process of evaluating and managing generated content to detect inappropriate content or the risk of copyright infringement.

[0898] "Notification" is the act of the server informing the user whether the generated content is appropriate or has been blocked because it is inappropriate.

[0899] "Natural Language Processing (NLP)" refers to the technology of analyzing, understanding, and appropriately processing human language. It is used for text analysis of generated content.

[0900] "Copyright infringement" is the act of using someone else's copyright without permission, which carries legal risks.

[0901] "Blocking" is the act of restricting access to prevent users from being provided with inappropriate content or generated content that poses a risk of copyright infringement.

[0902] This invention relates to a system that automatically filters and moderates content generated by generative AI, aiming to improve the quality of generated content and reduce legal risks. This system is composed of elements such as users, terminals, servers, and generative AI, and by clarifying the roles of each, it achieves efficient and safe content generation.

[0903] System Configuration

[0904] User

[0905] Users send content generation requests to the generative AI using a dedicated device, which can be a computer, smartphone, tablet, or other device, and access the generative AI model using a web browser or dedicated application.

[0906] Terminal

[0907] The terminal is a device through which users can send content generation requests to the generation AI, and acts as a user interface. A request form is displayed here, and users can enter the type of content they want to generate and guidelines.

[0908] server

[0909] The server receives user requests and forwards them appropriately to the AI ​​generator. It also receives the generated content returned by the AI ​​generator and automatically filters and moderates it. Specific software used in this process includes the natural language processing (NLP) libraries SpaCy and NLTK.

[0910] Generation AI

[0911] Generative AI is an artificial intelligence system that generates text content based on user requests, including language models like GPT-4.

[0912] Specific examples

[0913] 1. User Request Submission

[0914] A user sends a request from their device to "generate an adventure story for children." Specifically, they enter "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or pose a risk of copyright infringement" into the web form and click the submit button.

[0915] 2. Receiving and forwarding requests

[0916] The server receives the HTTP request, analyzes its contents, and converts it into JSON format for transmission to the API of the AI ​​generation. At this time, the request contents are transmitted to the endpoint of the AI ​​generation.

[0917] 3. Story Generation Using Generative AI

[0918] Based on the request, the AI ​​generates an "adventure story about a prince and princess meeting mysterious creatures in the forest." The generated content is sent back to the server in JSON format.

[0919] 4. Receiving and Processing Generated Content

[0920] The server receives the generated stories and stores them in an internal data structure, then uses an NLP library (e.g., SpaCy or NLTK) to analyze the text and check for profanity and similarity to existing works.

[0921] 5. Blocking inappropriate content

[0922] If the server determines that the content contains inappropriate content based on the analysis results, it will block the content. For example, if the content contains a scene in which the prince indiscriminately kills monsters, the filtering function will automatically detect it as inappropriate.

[0923] 6. Notice to Users

[0924] The server notifies the user whether the generated content is appropriate or has been blocked because it is inappropriate. Specifically, the user is shown a message saying, "The generation has been blocked because it contains inappropriate content."

[0925] Prompt Sentence Examples

[0926] Use the following prompt to input the generative AI model:

[0927] Create an adventure story for children. Please be careful not to include violent scenes or copyright infringement in the content.

[0928] This system allows users to use generative AI with confidence, reducing legal risks while producing high-quality content.

[0929] keyword

[0930] Generative AI model, prompt sentence

[0931] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0932] Step 1:

[0933] The user uses their device to send a content generation request to the generation AI. This request includes the type of content they want to generate and guidelines. Specifically, the user enters the following into a web form: "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or risk copyright infringement." and clicks the submit button. This form input is sent to the server as a request.

[0934] input:

[0935] User input into a web form (content generation request)

[0936] output:

[0937] HTTP request to the server

[0938] Step 2:

[0939] The server receives the request sent by the user. It analyzes the received request and converts it into a format suitable for the API of the generation AI (for example, JSON format). It then forwards the request to the endpoint of the generation AI. Specifically, the server checks the contents of the request, converts the format if necessary, and sends it to the API of the generation AI.

[0940] input:

[0941] HTTP request from the user

[0942] output:

[0943] The converted request to the generating AI (JSON format)

[0944] Step 3:

[0945] The generation AI generates content based on the received request. The generation AI uses an internal algorithm (e.g., GPT-4) to automatically generate text based on the input prompt. The generated content is sent back to the server in JSON format.

[0946] input:

[0947] The transformed request to the generative AI

[0948] output:

[0949] Generated content (JSON format)

[0950] Step 4:

[0951] The server receives the generated content from the generation AI, stores it in an internal data structure, and prepares it for the next filtering step. Specifically, the server parses the JSON-formatted data and extracts the text portion of the content.

[0952] input:

[0953] Generated content (JSON format)

[0954] output:

[0955] Storing text in internal data structures

[0956] Step 5:

[0957] The server automatically filters and moderates the generated content. During this process, it uses natural language processing (NLP) libraries (such as SpaCy or NLTK) to analyze the text and check for inappropriate content and similarities. Specifically, text analysis detects violent scenes and inappropriate content, and compares the generated content's similarities with existing databases. For example, if a video contains a scene in which a prince indiscriminately defeats monsters, it will be detected as violent.

[0958] input:

[0959] Generated content text

[0960] output:

[0961] Filtering and Moderation Results

[0962] Step 6:

[0963] If the server determines that the content is inappropriate or poses a risk of copyright infringement as a result of filtering and moderation, it will block the content. Specifically, if the filtering result is negative, the generated content will not be returned to the user and an error message will be prepared.

[0964] input:

[0965] Filtering and Moderation Results

[0966] output:

[0967] Generate appropriate content or error messages

[0968] Step 7:

[0969] The server notifies the user whether the generated content is appropriate or inappropriate and has been blocked. The server returns the generated content or an error message to the user's device as an HTTP response. Specifically, a message stating "Generation has been blocked because it contains inappropriate content" is displayed.

[0970] input:

[0971] Appropriate content or error messages

[0972] output:

[0973] HTTP response to the user

[0974] ---

[0975] The above is the specific flow of the program processing of this system. This allows users to use the generative AI with peace of mind, improves the quality of generated content, and reduces legal risks.

[0976] (Application example 1)

[0977] 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."

[0978] If the content generated by generative AI contains inappropriate content or poses a risk of copyright infringement, it is difficult to provide a safe environment for users. Furthermore, there is a lack of effective means to check for these risks before the generated content is distributed, raising concerns about legal risks and a decline in quality. In particular, in the case of educational content, the inclusion of inappropriate content not only reduces the educational effectiveness but also reduces the credibility of the content.

[0979] 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.

[0980] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying users of the generated content, means for checking the generated content before distribution, means for analyzing the audio and subtitles of the generated content using natural language processing, and means for evaluating the generated content based on guidelines specialized for educational content. This allows the generated content to be used and distributed safely without containing inappropriate content or the risk of copyright infringement.

[0981] "Generated content" refers to digital content generated by a generative AI in response to a user request.

[0982] "Filtering" refers to the process of automatically detecting and removing inappropriate language or content contained in generated content.

[0983] "Risk of copyright infringement" refers to situations in which generated content may infringe the rights of existing copyrighted works.

[0984] "Inappropriate content" refers to any expression or information that may violate user guidelines or the law.

[0985] "Notification to users" refers to a means of notifying users of the evaluation results and block information of generated content.

[0986] The "means for checking before distribution" refers to a process for confirming the appropriateness of generated content before distributing the content to a user or a third party.

[0987] "Natural language processing" refers to the technology that enables computers to understand, generate, and interpret human language.

[0988] "Means for analyzing audio and subtitles" refers to technology for automatically analyzing the audio data and subtitle data contained in the generated content and evaluating its content.

[0989] "Educational Content Specific Guidelines" means specific standards or rules that apply to generated content used for educational purposes.

[0990] This invention is a system for automatically filtering and moderating content generated by generative AI, with the aim of reducing the risk of inappropriate content and copyright infringement, particularly in the generation of educational content, by eliminating inappropriate content and providing reliable, high-quality content.

[0991] System Configuration

[0992] The system consists of the following main components:

[0993] 1. Server: Receives requests for generated content and forwards them to the generation AI. Receives the generated content and performs filtering and moderation. Hardware used could be an AWS EC2 instance, for example.

[0994] 2. Generative AI: Generates content based on user requests, for example using OpenAI's GPT-4.

[0995] 3. Natural language processing tools: Analyze the audio and subtitles of generated content to assess the risk of inappropriate content and copyright infringement. For example, Google Cloud Natural Language API is used.

[0996] 4. User device: Sends requests and receives generated content. Uses a smartphone (iOS or Android device).

[0997] Data processing and calculation

[0998] Receiving and forwarding requests: The server forwards requests for generated content received from the user device to the generation AI. The request includes the type of content to be generated and guidelines.

[0999] Content Generation: Generative AI generates content based on given guidelines, for example, given a prompt to generate an "educational video learning about the minerals of the Earth."

[1000] Filtering and moderation: The server automatically filters and moderates generated content, using natural language processing (Google Cloud Natural Language API) to analyze text and audio data and assess risk of inappropriate content and copyright infringement.

[1001] Notification to users: Based on the evaluation results, users will be notified whether the generated content is appropriate. If it is deemed inappropriate, users will be notified along with the reason.

[1002] Specific examples

[1003] A user might submit a request to generate educational video content, for example, "Please generate an educational video about the minerals of the Earth. Please do not include any inappropriate language or copyright infringing content."

[1004] The server forwards this request to the generation AI, which generates an educational video according to the specified guidelines. The server then receives the generated video and analyzes the video's audio and subtitles using natural language processing tools. If the analysis finds that the content does not comply with the educational guidelines or poses a risk of copyright infringement, the video is blocked. Finally, the server notifies the user of the appropriateness of the generated content.

[1005] In this way, users can confidently create and distribute high-quality, reliable educational content.

[1006] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1007] Step 1:

[1008] The user uses their device to send a request to the AI. The request includes the type of content they want to generate and guidelines. For example, they might enter a prompt like, "Generate an educational video that teaches about the minerals on Earth. Please do not include any inappropriate language or copyright infringement." This request is then sent to the server.

[1009] Step 2:

[1010] The server parses the request received from the user and converts it into an appropriate format. The server then forwards the request to the generation AI, which sends the request data to a specific API endpoint and applies prompts so that the generation AI can understand the content.

[1011] Step 3:

[1012] The generation AI generates content based on the request. The generation AI receives prompt text as input and outputs generated content such as an educational video script or narration. Each generation AI has different parameters and settings, so processing is done in the appropriate data format.

[1013] Step 4:

[1014] The server receives the generated content from the generation AI, records its contents, verifies the data format of the generated content (video script and narration), and converts it to fit the next processing step.

[1015] Step 5:

[1016] The server analyzes the generated content using natural language processing (NLP) tools. Audio and subtitle data is converted into text format and analyzed through a natural language processing engine (e.g., Google Cloud Natural Language API). The results of this analysis are used to assess the risk of inappropriate content and copyright infringement.

[1017] Step 6:

[1018] The server filters the generated content based on the analysis results. If inappropriate content or a risk of copyright infringement is detected, the server blocks the content. If necessary, the server modifies part of the generated content or requests the AI ​​to regenerate it.

[1019] Step 7:

[1020] The server notifies the user of the evaluation results of the generated content. If the content is appropriate, the server notifies the user and provides a download link and notification that the content is ready for distribution. If the content contains inappropriate content, the server notifies the user, along with the reason for the inappropriate content.

[1021] In this way, through a series of processing steps, a system is realized that ensures the quality of user-generated content and provides an environment in which it can be distributed safely.

[1022] 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.

[1023] This invention relates to a system that provides more appropriate content by combining a system that reduces the risk of inappropriate content and copyright infringement in content generated by generative AI with an emotion engine that recognizes user emotions. By taking user emotions into consideration, this system aims to further improve the quality of generated content and enhance the user experience.

[1024] System Overview

[1025] The system includes the following main functions:

[1026] 1. Automatically filtering generated content

[1027] 2. Means of assessing the risk of copyright infringement of generated content

[1028] 3. How to block generated content if it contains inappropriate content

[1029] 4. Means of notifying users about generated content

[1030] 5. Emotion engine that recognizes user emotions

[1031] 6. The ability to adjust generated content based on user sentiment

[1032] 7. Means for suspending provision of generated content based on emotion recognition results

[1033] Program processing

[1034] The program of this system performs processing in the following procedure.

[1035] 1. Activating the Emotional Engine

[1036] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time. This analysis recognizes the user's emotions.

[1037] 2. Submitting a content generation request

[1038] The user sends a content generation request to the AI ​​via their device, which includes the type of content they want to generate and guidelines.

[1039] 3. Receiving and forwarding requests

[1040] The server receives the user's request and forwards it to the generation AI.

[1041] 4. Content generation using generative AI

[1042] The generation AI generates content that meets the specified conditions based on the request.

[1043] 5. Receiving Generated Content

[1044] The server receives the generated content from the generation AI.

[1045] 6. Start filtering and moderation

[1046] Automatic filtering and moderation of server-generated content, using natural language processing (NLP) to analyze text and check for inappropriate language and copyright infringement.

[1047] 7. Emotion-aware content adjustment

[1048] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state and adjusts the content as necessary. For example, if the user is feeling stressed, the server may provide a relaxing story.

[1049] 8. Blocking inappropriate content

[1050] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[1051] 9. Notice to Users

[1052] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[1053] Specific examples

[1054] 1. User Emotion Recognition

[1055] While the user is viewing content on their mobile device, the emotion engine analyzes the user's facial expression data in real time and recognizes their "enjoyment" state.

[1056] 2. Sending a content generation request

[1057] A user submits a request to "generate an adventure story for children."

[1058] 3. Adventure Story Generation Using Generative AI

[1059] The AI ​​generates an "adventure story," which includes dangerous adventure scenes.

[1060] 4. Filtering and Moderation

[1061] The server performs filtering and moderation, determining that dangerous scenes are inappropriate and blocking them.

[1062] 5. Adjustment based on emotion recognition results

[1063] In order to keep the user in a "fun" state, the emotion engine adjusts to prioritize the generation of new stories based on "fun adventures."

[1064] 6. Delivery of Final Content

[1065] The server generates a tailored and appropriate adventure story and provides it to the user.

[1066] In this way, the system can provide appropriate content while taking into account the user's emotions, improving the overall user experience.

[1067] The processing flow will be explained below.

[1068] Specific processing flow of the program

[1069] Step 1:

[1070] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time, thereby recognizing the user's emotional state.

[1071] Step 2:

[1072] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[1073] Step 3:

[1074] The server receives the user's request, analyzes the request content, and prepares to forward it to the generation AI in an appropriate format.

[1075] Step 4:

[1076] The server forwards a content generation request to the generation AI, and the generation AI starts generating content based on the specified conditions.

[1077] Step 5:

[1078] The generation AI generates the specified content based on the request, for example, "A story about a prince and princess having an adventure in the forest."

[1079] Step 6:

[1080] The server receives the generated content from the generative AI, which has not yet been filtered or moderated.

[1081] Step 7:

[1082] The server analyzes the generated content using natural language processing (NLP) to check for inappropriate words and expressions.

[1083] Step 8:

[1084] The server evaluates the risk of copyright infringement of generated content by matching the content similarity with existing databases. If similar content is found, the content is determined to be at risk of copyright infringement.

[1085] Step 9:

[1086] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state. For example, if the user is feeling anxious, the content can be adjusted to change it into a relaxing story.

[1087] Step 10:

[1088] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters is deemed inappropriate and blocked.

[1089] Step 11:

[1090] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[1091] Step 12:

[1092] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they can use it, and if it is inappropriate, they are given the option to send the request again or receive the adjusted content.

[1093] Specific examples

[1094] 1. Step 1-6: The user sends a request to "generate an adventure story for children," and the server receives the request and forwards it to the generation AI. The generation AI generates a story about a prince and princess having an adventure in the forest, and the server receives it.

[1095] 2. Step 7: The server analyzes the generated story using natural language processing, for example, to check whether it contains any "dangerous scenes."

[1096] 3. Step 8: The server checks the generated content against an existing database to assess the risk of copyright infringement.

[1097] 4. Step 9: The emotion engine recognizes the user's "anxiety" state, and the server adjusts the story content to a "relaxing adventure story" based on this result.

[1098] 5. Steps 10-12: The server determines that the final adjusted story is appropriate and notifies the user, who then accepts and views the story.

[1099] In this way, the system automatically checks the quality and legal risks of generated content, and further adjusts the content based on user sentiment, providing an environment where users can use it with peace of mind.

[1100] Example 2

[1101] 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."

[1102] The content generated by modern generative AI models may contain inappropriate content or risk copyright infringement. Furthermore, if the generated content does not match the user's emotional state, it can negatively impact the user experience. To address this issue, it is necessary to ensure the appropriateness of the content and take the user's emotional state into consideration.

[1103] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying the user about the generated content, means for recognizing the user's emotion, means for adjusting the content of the generated content based on the user's emotion, and means for suspending provision of the generated content based on the result of the user's emotion recognition. This makes it possible to provide content that matches the user's emotional state while ensuring the appropriateness of the generated content.

[1104] "Generated content" refers to information such as text, images, and video that is automatically generated by a generative AI model based on a user request.

[1105] "Filtering measures" are technologies or algorithms used to analyze generated content and determine whether it contains inappropriate content or is at risk of copyright infringement.

[1106] "Copyright infringement risk assessment means" means any technology or algorithm that compares generated content with existing databases or copyrighted material to assess the risk of copyright infringement.

[1107] "Blocking measures" are technologies or methods that remove generated content before it is provided to users if the content contains inappropriate content or poses a risk of copyright infringement.

[1108] A "notification mechanism" is a mechanism for informing users of the outcome of generated content, whether it is appropriate or blocked for being inappropriate.

[1109] "Emotion recognition means" refers to technology or algorithms that analyze a user's facial expressions, voice, and input data, and recognize the user's emotions in real time based on the results.

[1110] The "emotion-based content adjustment means" refers to a technology or algorithm for changing the content of generated content to match the emotional state of the user based on the result of the user's emotion recognition.

[1111] The "provision interruption means" is a technique or method for interrupting the provision of generated content midway when the user's emotion recognition result satisfies certain conditions.

[1112] MODE FOR CARRYING OUT THE INVENTION

[1113] This invention is a system that reduces the risk of inappropriate content and copyright infringement in content generated using a generative AI model, and combines it with an emotion engine that recognizes user emotions to provide more appropriate content. This system aims to further improve the quality of generated content and enhance the user experience by taking user emotions into consideration.

[1114] Hardware and software used

[1115] This system mainly uses the following hardware and software:

[1116] Server: A high-performance server for data analysis and processing. Specific examples include server instances on AWS or Google Cloud Platform.

[1117] Terminal: A device operated by a user. Examples include PCs, smartphones, tablets, etc.

[1118] Emotion engine: Software for analyzing a user's facial expressions and voice data. Examples include OpenCV and Microsoft Azure Emotion API.

[1119] Generative AI model: An artificial intelligence model for content generation. An example of this is OpenAI GPT-3.

[1120] Natural Language Processing (NLP) software, used to analyze generated content and check for inappropriate content and copyright infringement risks. Examples include spaCy and NLTK.

[1121] Specific operation of the system

[1122] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expressions and voice data through the connected camera and microphone. The collected data is analyzed in real time to recognize the user's emotions.

[1123] The user fills out a content generation request form for the AI ​​on their device screen and clicks the "Submit" button. The request includes the type of content they want to generate and guidelines.

[1124] The server receives content generation requests sent by users, formats them as necessary, and forwards them to the generative AI model, which generates content based on the requests and meets the specified conditions.

[1125] The server receives the content generated by the generative AI model, analyzes it using NLP software, and blocks it if it contains inappropriate content or is at risk of copyright infringement.

[1126] The emotion engine analyzes the user's emotional state and determines whether the content is appropriate for the user's state. If necessary, it adjusts the generated content. For example, if the user is feeling stressed, it changes the content to something more relaxing.

[1127] Finally, the server checks the generated content for appropriateness and provides it to the user. If the content is inappropriate, the server notifies the user of the result.

[1128] Specific examples

[1129] As an example, consider the case where a user submits a content generation request using the following prompt:

[1130] Example prompt sentence:

[1131] "Generate fun adventure stories for children, but please do not include dangerous or violent scenes."

[1132] Based on this request, a generative AI model generates an adventure story. The server analyzes the content to ensure it does not contain any dangerous scenes. Furthermore, an emotion engine analyzes the user's emotional state and adjusts the content as needed. The optimal content is then delivered to the user.

[1133] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1134] Step 1:

[1135] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expression and voice data through the connected camera and microphone and analyzes it in real time. The input data is the user's facial expression and voice data, and the output is the analysis result that indicates the user's emotional state. Specifically, it calls the emotion engine's API and analyzes the collected data.

[1136] Step 2:

[1137] The user fills out a content generation request form for the generation AI on their device screen and clicks the "Submit" button. The input data is form data that includes the type of content they want to generate and guidelines, and the output data is the request data that is sent to the server. Specifically, the user fills out the form on the browser screen and sends a POST request.

[1138] Step 3:

[1139] The server receives a content generation request sent by the user. The received data is the user's request data, and the output is formatted request data to be sent to the generative AI model. The server formats the received request content as needed and sends it to the generative AI model. Specifically, it converts the received data into an internal format and sends it to the generative AI's API.

[1140] Step 4:

[1141] The generation AI generates content that meets the specified conditions based on the request. The input is the request data transferred from the server, and the output is the generated content data. Specifically, the generation AI runs a natural language generation model (e.g., GPT-3) based on the prompt sentence and outputs the generated results as text data.

[1142] Step 5:

[1143] The server receives the content generated by the generation AI. The input is the content data from the generation AI, and the output is the data to be analyzed for filtering and moderation. Specifically, the server receives the HTTP response and stores the generated content data in an internal database.

[1144] Step 6:

[1145] The server automatically filters and moderates the content it receives. The input is the content data to be analyzed, and the output is the filtering and moderation results. Specifically, it uses an NLP library (e.g., spaCy or NLTK) to analyze the text and check for inappropriate content and copyright infringement risks.

[1146] Step 7:

[1147] Based on the results of the emotion engine, the server determines whether the content is appropriate for the user's emotional state and adjusts the content as necessary. The input is the emotion recognition result and the content data to be analyzed, and the output is the adjusted content data. Specifically, the emotion recognition result is obtained from the API, and the content is regenerated or replaced with a template as necessary.

[1148] Step 8:

[1149] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it blocks the content. The input is the filtering and moderation results, and the output is a list of blocked content. The specific operation is to add the content to the block list so that it is not provided to users.

[1150] Step 9:

[1151] The server notifies the user whether the generated content is appropriate or blocked because it is inappropriate. The input is the user's notification destination information and notification content data, and the output is the notification sending result. Specifically, the server notifies the user of the result via email or an in-app notification system.

[1152] (Application example 2)

[1153] 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."

[1154] Content generated by conventional generative AI models is not optimized based on user sentiment, and therefore does not provide a sufficiently good user experience. Furthermore, content may contain inappropriate content or risk copyright infringement, making it difficult to provide safe and appropriate content. This has led to problems such as reduced user satisfaction.

[1155] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1156] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking generated content if it contains inappropriate content, means for notifying users about the generated content, means for analyzing emotions of users, and means for adjusting the content of the generated content based on the emotions of users, thereby enabling the provision of appropriate content that takes into consideration the emotions of users.

[1157] "Means for automatically filtering generated content" refers to technology that automatically analyzes generated content and detects and removes inappropriate content or harmful information.

[1158] The "means for assessing the risk of copyright infringement of generated content" is a technology for assessing whether generated content uses existing copyrighted works without permission and determining the risk of copyright infringement.

[1159] "Means for blocking generated content if it contains inappropriate content" refers to technology that removes or stops the display of generated content if it contains inappropriate language or harmful information.

[1160] The "means for notifying users of generated content" is a technique for notifying users whether generated content is in an appropriate state when provided to the users.

[1161] "Emotion analysis means for recognizing user emotions" is a technology that analyzes the user's facial expressions, tone of voice, input data, etc. to recognize the user's emotional state in real time.

[1162] The "means for adjusting the content of generated content based on the user's emotions" is a technology for optimizing the content of generated content in accordance with the recognized user's emotions.

[1163] System Overview

[1164] The system for implementing this invention integrates multiple key functions, analyzes user emotions in real time, and appropriately filters and adjusts the generated content to provide high-quality content.

[1165] Program processing

[1166] The program of this system includes the following elements:

[1167] 1. Starting and analyzing the emotion recognition engine

[1168] The server collects data in real time from the camera and microphone of the user's device (smartphone, tablet, etc.), which allows it to analyze the user's facial expressions and voice and recognize their emotions.

[1169] The software used is a general image analysis API (e.g., image analysis cloud service) for image analysis, and an audio tone analysis tool (e.g., audio tone analysis service) for audio analysis.

[1170] 2. Submitting a content generation request

[1171] The user sends a request to the AI ​​via their device, including the type of content they want and keywords, and the request includes a specific prompt.

[1172] 3. Content generation using generative AI models

[1173] The server generates content based on specified conditions based on a generative AI model (e.g., a generative AI service).

[1174] The generated content is diverse, including text, images, and audio.

[1175] Specific examples

[1176] For example, if a user wants to send a request to create relaxing music, they might use a prompt like this:

[1177] Example prompt: "Generate relaxing music containing nature sounds that users would like to listen to while in a relaxed state."

[1178] Based on this request, the server issues instructions to the generation AI to generate the requested content.

[1179] 4. Filtering and Moderation

[1180] The server receives the generated content and automatically checks it for inappropriate content and risks of copyright infringement.

[1181] It utilizes natural language processing (NLP) to analyze text and uses machine learning libraries such as TensorFlow and PyTorch.

[1182] 5. Emotion-based content adjustment

[1183] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, the server regenerates content to provide soothing sensations without overstimulating them.

[1184] The parameters of the AI ​​model are adjusted based on the results of sentiment analysis to generate new content.

[1185] 6. Provision of Final Content and Notification

[1186] The server provides the tailored content to the user and notifies them if the content is appropriate.

[1187] If the user is in a relaxed state, content with a high relaxation effect is provided, improving the user experience.

[1188] As described above, the present invention takes into consideration the feelings of users and makes it possible to provide optimal content while reducing the risk of inappropriate content and copyright infringement.

[1189] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1190] Step 1:

[1191] The server collects data in real time from the camera and microphone of the device (smartphone, tablet, etc.). This data is used to obtain the user's facial expressions and voice data, and preprocesses it to recognize emotions. The collected data is analyzed using an image analysis cloud service and a voice tone analysis service, and the user's emotional state (e.g., "relaxed") is output.

[1192] Step 2:

[1193] The user inputs and sends a request through their device, including the type of content they want and keywords. The generated prompt is something like "Generate relaxing music." The server receives this request and prepares the prompt as data to send appropriate instructions to the generative AI model.

[1194] Step 3:

[1195] The server sends a prompt to the generative AI model, instructing it to generate content based on the specified conditions. The generative AI model generates content such as text, images, and audio based on the request. This generated content is then sent back to the server.

[1196] Step 4:

[1197] The server receives the generated content and automatically filters it for inappropriate content and copyright infringement risks. It uses natural language processing (NLP) to analyze the text and machine learning libraries (e.g., TensorFlow and PyTorch) to assess the risk. Based on this assessment, it blocks the content if it contains inappropriate content.

[1198] Step 5:

[1199] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, it checks whether the generated content is overly stimulating and regenerates it if necessary. It also adjusts the parameters of the AI ​​model to optimize the newly generated content.

[1200] Step 6:

[1201] The server provides the final adjusted content to the user and notifies them whether the content is appropriate. As the content is provided, the emotion engine analyzes it again in light of the user's emotions and provides appropriate feedback. If the user is relaxed, relaxation content is provided to maintain that state.

[1202] As described above, this system specifically realizes a mechanism for analyzing user emotions in real time and generating and providing optimal content based on that.

[1203] 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.

[1204] 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.

[1205] 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.

[1206] [Fourth embodiment]

[1207] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1208] 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.

[1209] 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).

[1210] 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.

[1211] 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.

[1212] 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).

[1213] 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.

[1214] 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.

[1215] 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.

[1216] 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.

[1217] 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.

[1218] 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.

[1219] 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."

[1220] The present invention relates to a system for automatically filtering and moderating content generated by generative AI to reduce the risk of inappropriate content and copyright infringement. The system provides a means for analyzing the generated content and detecting the risk of inappropriate content and copyright infringement.

[1221] System Overview

[1222] The system includes the following main functions:

[1223] 1. Automatically filtering generated content

[1224] 2. Means of assessing the risk of copyright infringement of generated content

[1225] 3. How to block generated content if it contains inappropriate content

[1226] 4. Means of notifying users about generated content

[1227] Program processing

[1228] The program of this system performs processing in the following procedure.

[1229] 1. Submitting a content generation request

[1230] The user uses a device to send a content generation request to the generation AI, which includes the type of content they want to generate and guidelines.

[1231] 2. Receiving and forwarding requests

[1232] The server receives the request from the user and forwards the request to the generation AI.

[1233] 3. Content generation using generative AI

[1234] The generation AI generates the specified content based on the request.

[1235] 4. Receiving Generated Content

[1236] The server receives the generated content from the generation AI.

[1237] 5. Start filtering and moderation

[1238] Server-generated content is automatically filtered and moderated, using natural language processing (NLP) to analyze text and detect inappropriate language, and it also compares content similarity with existing databases to assess the risk of copyright infringement.

[1239] 6. Blocking inappropriate content

[1240] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[1241] 7. Notice to Users

[1242] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[1243] Specific examples

[1244] 1. User Request Submission

[1245] A user submits a request to "generate an adventure story for children."

[1246] 2. Receiving and forwarding requests

[1247] The server receives the request and forwards it to the generation AI.

[1248] 3. Story Generation Using Generative AI

[1249] The AI ​​generates an adventure story in which a prince and princess encounter mysterious creatures in the forest.

[1250] 4. Receiving Generated Content

[1251] The server receives the generated story.

[1252] 5. Filtering and moderation processes

[1253] The server analyzes the story text and checks for inappropriate language and similarities to existing works. For example, if a story contains a scene in which the prince indiscriminately defeats monsters, it will automatically detect it as inappropriate.

[1254] 6. Blocking inappropriate content

[1255] The server detects inappropriate content and blocks the story.

[1256] 7. Notifying Users and Providing Results

[1257] The server sends a notification to the user saying "Generation blocked due to inappropriate content."

[1258] As a result, the present invention provides an environment in which users can use generative AI with peace of mind, improving the quality of generated content and reducing legal risks.

[1259] The processing flow will be explained below.

[1260] Program processing flow

[1261] Step 1:

[1262] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[1263] Step 2:

[1264] The server receives the user's request, which includes details about the content, and prepares to parse it and forward it to the generation AI in the appropriate format.

[1265] Step 3:

[1266] The server forwards the content generation request to the generation AI, which starts the process of the generation AI generating content based on the specified conditions.

[1267] Step 4:

[1268] Generative AI generates content based on user requests, such as "A story about a prince and princess having an adventure in the forest."

[1269] Step 5:

[1270] The server receives the generated content from the generation AI. This content is raw, as it has not yet been filtered or moderated.

[1271] Step 6:

[1272] The server then begins the filtering and moderation process on the generated content, first using natural language processing (NLP) to analyze the text and check for inappropriate language or expressions.

[1273] Step 7:

[1274] The server then compares the content similarity with existing databases to assess the risk of copyright infringement of the generated content. If similar content is found, the content is deemed to be at risk of copyright infringement.

[1275] Step 8:

[1276] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters would be deemed inappropriate.

[1277] Step 9:

[1278] The server generates a result indicating whether the content is appropriate or not and notifies the user. If the content is appropriate, it is served to the user as is, otherwise an error message is sent.

[1279] Step 10:

[1280] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they use it, and if it is inappropriate, they send a request again.

[1281] In this way, the system automatically checks the quality and legal risks of generated content, providing a safe and secure environment for users.

[1282] Example 1

[1283] 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."

[1284] Since content generated using generative AI may contain inappropriate content or may infringe copyrights, there is no environment in place for users to use it safely. For this reason, there is a need for a system that can automatically filter generated content and provide only appropriate content.

[1285] 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.

[1286] In this invention, the server includes: a means for a user to send a content generation request to the generation AI using a terminal; a means for the server to receive the request from the user and forward it to the generation AI; a means for the generation AI to generate content based on the request; a means for the server to receive the generated content from the generation AI; a means for the server to automatically filter and moderate the generated content; a means for the server to block the generated content if the server determines that the generated content contains inappropriate content or poses a risk of copyright infringement; and a means for the server to notify the user of the result of whether the generated content is appropriate or has been blocked. This makes it possible to improve the quality of generated content and reduce legal risks.

[1287] "User" means an individual or entity that uses a device to send a content generation request to a generation AI.

[1288] A "terminal" is an electronic device used by a user to send a content generation request to the generation AI, including a computer, smartphone, tablet, etc.

[1289] "Generative AI" is an artificial intelligence system that automatically generates content based on a given request.

[1290] A "server" is a computer system that receives requests from users via the Internet, forwards them to the generation AI, and receives and processes the generated content.

[1291] A "content generation request" is an instruction sent by a user to a generation AI to generate specific content, including the content and guidelines.

[1292] "Filtering" is the process of automatically detecting and filtering out inappropriate content from generated content.

[1293] "Moderation" is the process of evaluating and managing generated content to detect inappropriate content or the risk of copyright infringement.

[1294] "Notification" is the act of the server informing the user whether the generated content is appropriate or has been blocked because it is inappropriate.

[1295] "Natural Language Processing (NLP)" refers to the technology of analyzing, understanding, and appropriately processing human language. It is used for text analysis of generated content.

[1296] "Copyright infringement" is the act of using someone else's copyright without permission, which carries legal risks.

[1297] "Blocking" is the act of restricting access to prevent users from being provided with inappropriate content or generated content that poses a risk of copyright infringement.

[1298] This invention relates to a system that automatically filters and moderates content generated by generative AI, aiming to improve the quality of generated content and reduce legal risks. This system is composed of elements such as users, terminals, servers, and generative AI, and by clarifying the roles of each, it achieves efficient and safe content generation.

[1299] System Configuration

[1300] User

[1301] Users send content generation requests to the generative AI using a dedicated device, which can be a computer, smartphone, tablet, or other device, and access the generative AI model using a web browser or dedicated application.

[1302] Terminal

[1303] The terminal is a device through which users can send content generation requests to the generation AI, and acts as a user interface. A request form is displayed here, and users can enter the type of content they want to generate and guidelines.

[1304] server

[1305] The server receives user requests and forwards them appropriately to the AI ​​generator. It also receives the generated content returned by the AI ​​generator and automatically filters and moderates it. Specific software used in this process includes the natural language processing (NLP) libraries SpaCy and NLTK.

[1306] Generation AI

[1307] Generative AI is an artificial intelligence system that generates text content based on user requests, including language models like GPT-4.

[1308] Specific examples

[1309] 1. User Request Submission

[1310] A user sends a request from their device to "generate an adventure story for children." Specifically, they enter "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or pose a risk of copyright infringement" into the web form and click the submit button.

[1311] 2. Receiving and forwarding requests

[1312] The server receives the HTTP request, analyzes its contents, and converts it into JSON format for transmission to the API of the AI ​​generation. At this time, the request contents are transmitted to the endpoint of the AI ​​generation.

[1313] 3. Story Generation Using Generative AI

[1314] Based on the request, the AI ​​generates an "adventure story about a prince and princess meeting mysterious creatures in the forest." The generated content is sent back to the server in JSON format.

[1315] 4. Receiving and Processing Generated Content

[1316] The server receives the generated stories and stores them in an internal data structure, then uses an NLP library (e.g., SpaCy or NLTK) to analyze the text and check for profanity and similarity to existing works.

[1317] 5. Blocking inappropriate content

[1318] If the server determines that the content contains inappropriate content based on the analysis results, it will block the content. For example, if the content contains a scene in which the prince indiscriminately kills monsters, the filtering function will automatically detect it as inappropriate.

[1319] 6. Notice to Users

[1320] The server notifies the user whether the generated content is appropriate or has been blocked because it is inappropriate. Specifically, the user is shown a message saying, "The generation has been blocked because it contains inappropriate content."

[1321] Prompt Sentence Examples

[1322] Use the following prompt to input the generative AI model:

[1323] Create an adventure story for children. Please be careful not to include violent scenes or copyright infringement in the content.

[1324] This system allows users to use generative AI with confidence, reducing legal risks while producing high-quality content.

[1325] keyword

[1326] Generative AI model, prompt sentence

[1327] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1328] Step 1:

[1329] The user uses their device to send a content generation request to the generation AI. This request includes the type of content they want to generate and guidelines. Specifically, the user enters the following into a web form: "Please generate an adventure story for children. Please be careful that the content does not contain violent scenes or risk copyright infringement." and clicks the submit button. This form input is sent to the server as a request.

[1330] input:

[1331] User input into a web form (content generation request)

[1332] output:

[1333] HTTP request to the server

[1334] Step 2:

[1335] The server receives the request sent by the user. It analyzes the received request and converts it into a format suitable for the API of the generation AI (for example, JSON format). It then forwards the request to the endpoint of the generation AI. Specifically, the server checks the contents of the request, converts the format if necessary, and sends it to the API of the generation AI.

[1336] input:

[1337] HTTP request from the user

[1338] output:

[1339] The converted request to the generating AI (JSON format)

[1340] Step 3:

[1341] The generation AI generates content based on the received request. The generation AI uses an internal algorithm (e.g., GPT-4) to automatically generate text based on the input prompt. The generated content is sent back to the server in JSON format.

[1342] input:

[1343] The transformed request to the generative AI

[1344] output:

[1345] Generated content (JSON format)

[1346] Step 4:

[1347] The server receives the generated content from the generation AI, stores it in an internal data structure, and prepares it for the next filtering step. Specifically, the server parses the JSON-formatted data and extracts the text portion of the content.

[1348] input:

[1349] Generated content (JSON format)

[1350] output:

[1351] Storing text in internal data structures

[1352] Step 5:

[1353] The server automatically filters and moderates the generated content. During this process, it uses natural language processing (NLP) libraries (such as SpaCy or NLTK) to analyze the text and check for inappropriate content and similarities. Specifically, text analysis detects violent scenes and inappropriate content, and compares the generated content's similarities with existing databases. For example, if a video contains a scene in which a prince indiscriminately defeats monsters, it will be detected as violent.

[1354] input:

[1355] Generated content text

[1356] output:

[1357] Filtering and Moderation Results

[1358] Step 6:

[1359] If the server determines that the content is inappropriate or poses a risk of copyright infringement as a result of filtering and moderation, it will block the content. Specifically, if the filtering result is negative, the generated content will not be returned to the user and an error message will be prepared.

[1360] input:

[1361] Filtering and Moderation Results

[1362] output:

[1363] Generate appropriate content or error messages

[1364] Step 7:

[1365] The server notifies the user whether the generated content is appropriate or inappropriate and has been blocked. The server returns the generated content or an error message to the user's device as an HTTP response. Specifically, a message stating "Generation has been blocked because it contains inappropriate content" is displayed.

[1366] input:

[1367] Appropriate content or error messages

[1368] output:

[1369] HTTP response to the user

[1370] ---

[1371] The above is the specific flow of the program processing of this system. This allows users to use the generative AI with peace of mind, improves the quality of generated content, and reduces legal risks.

[1372] (Application example 1)

[1373] 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."

[1374] If the content generated by generative AI contains inappropriate content or poses a risk of copyright infringement, it is difficult to provide a safe environment for users. Furthermore, there is a lack of effective means to check for these risks before the generated content is distributed, raising concerns about legal risks and a decline in quality. In particular, in the case of educational content, the inclusion of inappropriate content not only reduces the educational effectiveness but also reduces the credibility of the content.

[1375] 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.

[1376] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying users of the generated content, means for checking the generated content before distribution, means for analyzing the audio and subtitles of the generated content using natural language processing, and means for evaluating the generated content based on guidelines specialized for educational content. This allows the generated content to be used and distributed safely without containing inappropriate content or the risk of copyright infringement.

[1377] "Generated content" refers to digital content generated by a generative AI in response to a user request.

[1378] "Filtering" refers to the process of automatically detecting and removing inappropriate language or content contained in generated content.

[1379] "Risk of copyright infringement" refers to situations in which generated content may infringe the rights of existing copyrighted works.

[1380] "Inappropriate content" refers to any expression or information that may violate user guidelines or the law.

[1381] "Notification to users" refers to a means of notifying users of the evaluation results and block information of generated content.

[1382] The "means for checking before distribution" refers to a process for confirming the appropriateness of generated content before distributing the content to a user or a third party.

[1383] "Natural language processing" refers to the technology that enables computers to understand, generate, and interpret human language.

[1384] "Means for analyzing audio and subtitles" refers to technology for automatically analyzing the audio data and subtitle data contained in the generated content and evaluating its content.

[1385] "Educational Content Specific Guidelines" means specific standards or rules that apply to generated content used for educational purposes.

[1386] This invention is a system for automatically filtering and moderating content generated by generative AI, with the aim of reducing the risk of inappropriate content and copyright infringement, particularly in the generation of educational content, by eliminating inappropriate content and providing reliable, high-quality content.

[1387] System Configuration

[1388] The system consists of the following main components:

[1389] 1. Server: Receives requests for generated content and forwards them to the generation AI. Receives the generated content and performs filtering and moderation. Hardware used could be an AWS EC2 instance, for example.

[1390] 2. Generative AI: Generates content based on user requests, for example using OpenAI's GPT-4.

[1391] 3. Natural language processing tools: Analyze the audio and subtitles of generated content to assess the risk of inappropriate content and copyright infringement. For example, Google Cloud Natural Language API is used.

[1392] 4. User device: Sends requests and receives generated content. Uses a smartphone (iOS or Android device).

[1393] Data processing and calculation

[1394] Receiving and forwarding requests: The server forwards requests for generated content received from the user device to the generation AI. The request includes the type of content to be generated and guidelines.

[1395] Content Generation: Generative AI generates content based on given guidelines, for example, given a prompt to generate an "educational video learning about the minerals of the Earth."

[1396] Filtering and moderation: The server automatically filters and moderates generated content, using natural language processing (Google Cloud Natural Language API) to analyze text and audio data and assess risk of inappropriate content and copyright infringement.

[1397] Notification to users: Based on the evaluation results, users will be notified whether the generated content is appropriate. If it is deemed inappropriate, users will be notified along with the reason.

[1398] Specific examples

[1399] A user might submit a request to generate educational video content, for example, "Please generate an educational video about the minerals of the Earth. Please do not include any inappropriate language or copyright infringing content."

[1400] The server forwards this request to the generation AI, which generates an educational video according to the specified guidelines. The server then receives the generated video and analyzes the video's audio and subtitles using natural language processing tools. If the analysis finds that the content does not comply with the educational guidelines or poses a risk of copyright infringement, the video is blocked. Finally, the server notifies the user of the appropriateness of the generated content.

[1401] In this way, users can confidently create and distribute high-quality, reliable educational content.

[1402] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1403] Step 1:

[1404] The user uses their device to send a request to the AI. The request includes the type of content they want to generate and guidelines. For example, they might enter a prompt like, "Generate an educational video that teaches about the minerals on Earth. Please do not include any inappropriate language or copyright infringement." This request is then sent to the server.

[1405] Step 2:

[1406] The server parses the request received from the user and converts it into an appropriate format. The server then forwards the request to the generation AI, which sends the request data to a specific API endpoint and applies prompts so that the generation AI can understand the content.

[1407] Step 3:

[1408] The generation AI generates content based on the request. The generation AI receives prompt text as input and outputs generated content such as an educational video script or narration. Each generation AI has different parameters and settings, so processing is done in the appropriate data format.

[1409] Step 4:

[1410] The server receives the generated content from the generation AI, records its contents, verifies the data format of the generated content (video script and narration), and converts it to fit the next processing step.

[1411] Step 5:

[1412] The server analyzes the generated content using natural language processing (NLP) tools. Audio and subtitle data is converted into text format and analyzed through a natural language processing engine (e.g., Google Cloud Natural Language API). The results of this analysis are used to assess the risk of inappropriate content and copyright infringement.

[1413] Step 6:

[1414] The server filters the generated content based on the analysis results. If inappropriate content or a risk of copyright infringement is detected, the server blocks the content. If necessary, the server modifies part of the generated content or requests the AI ​​to regenerate it.

[1415] Step 7:

[1416] The server notifies the user of the evaluation results of the generated content. If the content is appropriate, the server notifies the user and provides a download link and notification that the content is ready for distribution. If the content contains inappropriate content, the server notifies the user, along with the reason for the inappropriate content.

[1417] In this way, through a series of processing steps, a system is realized that ensures the quality of user-generated content and provides an environment in which it can be distributed safely.

[1418] 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.

[1419] This invention relates to a system that provides more appropriate content by combining a system that reduces the risk of inappropriate content and copyright infringement in content generated by generative AI with an emotion engine that recognizes user emotions. By taking user emotions into consideration, this system aims to further improve the quality of generated content and enhance the user experience.

[1420] System Overview

[1421] The system includes the following main functions:

[1422] 1. Automatically filtering generated content

[1423] 2. Means of assessing the risk of copyright infringement of generated content

[1424] 3. How to block generated content if it contains inappropriate content

[1425] 4. Means of notifying users about generated content

[1426] 5. Emotion engine that recognizes user emotions

[1427] 6. The ability to adjust generated content based on user sentiment

[1428] 7. Means for suspending provision of generated content based on emotion recognition results

[1429] Program processing

[1430] The program of this system performs processing in the following procedure.

[1431] 1. Activating the Emotional Engine

[1432] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time. This analysis recognizes the user's emotions.

[1433] 2. Submitting a content generation request

[1434] The user sends a content generation request to the AI ​​via their device, which includes the type of content they want to generate and guidelines.

[1435] 3. Receiving and forwarding requests

[1436] The server receives the user's request and forwards it to the generation AI.

[1437] 4. Content generation using generative AI

[1438] The generation AI generates content that meets the specified conditions based on the request.

[1439] 5. Receiving Generated Content

[1440] The server receives the generated content from the generation AI.

[1441] 6. Start filtering and moderation

[1442] Automatic filtering and moderation of server-generated content, using natural language processing (NLP) to analyze text and check for inappropriate language and copyright infringement.

[1443] 7. Emotion-aware content adjustment

[1444] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state and adjusts the content as necessary. For example, if the user is feeling stressed, the server may provide a relaxing story.

[1445] 8. Blocking inappropriate content

[1446] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it will block the content.

[1447] 9. Notice to Users

[1448] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[1449] Specific examples

[1450] 1. User Emotion Recognition

[1451] While the user is viewing content on their mobile device, the emotion engine analyzes the user's facial expression data in real time and recognizes their "enjoyment" state.

[1452] 2. Sending a content generation request

[1453] A user submits a request to "generate an adventure story for children."

[1454] 3. Adventure Story Generation Using Generative AI

[1455] The AI ​​generates an "adventure story," which includes dangerous adventure scenes.

[1456] 4. Filtering and Moderation

[1457] The server performs filtering and moderation, determining that dangerous scenes are inappropriate and blocking them.

[1458] 5. Adjustment based on emotion recognition results

[1459] In order to keep the user in a "fun" state, the emotion engine adjusts to prioritize the generation of new stories based on "fun adventures."

[1460] 6. Delivery of Final Content

[1461] The server generates a tailored and appropriate adventure story and provides it to the user.

[1462] In this way, the system can provide appropriate content while taking into account the user's emotions, improving the overall user experience.

[1463] The processing flow will be explained below.

[1464] Specific processing flow of the program

[1465] Step 1:

[1466] The server activates the emotion engine and analyzes the user's facial expressions, voice, input data, etc. in real time, thereby recognizing the user's emotional state.

[1467] Step 2:

[1468] The user uses a device to send a content generation request to the AI, which includes the type of content they want to generate and guidelines.

[1469] Step 3:

[1470] The server receives the user's request, analyzes the request content, and prepares to forward it to the generation AI in an appropriate format.

[1471] Step 4:

[1472] The server forwards a content generation request to the generation AI, and the generation AI starts generating content based on the specified conditions.

[1473] Step 5:

[1474] The generation AI generates the specified content based on the request, for example, "A story about a prince and princess having an adventure in the forest."

[1475] Step 6:

[1476] The server receives the generated content from the generative AI, which has not yet been filtered or moderated.

[1477] Step 7:

[1478] The server analyzes the generated content using natural language processing (NLP) to check for inappropriate words and expressions.

[1479] Step 8:

[1480] The server evaluates the risk of copyright infringement of generated content by matching the content similarity with existing databases. If similar content is found, the content is determined to be at risk of copyright infringement.

[1481] Step 9:

[1482] The server uses the results of the emotion engine to determine whether the content is appropriate for the user's emotional state. For example, if the user is feeling anxious, the content can be adjusted to change it into a relaxing story.

[1483] Step 10:

[1484] The server evaluates the results of filtering and moderation, and blocks content that is deemed inappropriate or poses a risk of copyright infringement. For example, a scene in which a prince indiscriminately defeats monsters is deemed inappropriate and blocked.

[1485] Step 11:

[1486] The server notifies the user whether the generated content is appropriate or whether it is inappropriate and therefore blocked.

[1487] Step 12:

[1488] The user receives a notification from the server via their device and checks the generated content. If the content is appropriate, they can use it, and if it is inappropriate, they are given the option to send the request again or receive the adjusted content.

[1489] Specific examples

[1490] 1. Step 1-6: The user sends a request to "generate an adventure story for children," and the server receives the request and forwards it to the generation AI. The generation AI generates a story about a prince and princess having an adventure in the forest, and the server receives it.

[1491] 2. Step 7: The server analyzes the generated story using natural language processing, for example, to check whether it contains any "dangerous scenes."

[1492] 3. Step 8: The server checks the generated content against an existing database to assess the risk of copyright infringement.

[1493] 4. Step 9: The emotion engine recognizes the user's "anxiety" state, and the server adjusts the story content to a "relaxing adventure story" based on this result.

[1494] 5. Steps 10-12: The server determines that the final adjusted story is appropriate and notifies the user, who then accepts and views the story.

[1495] In this way, the system automatically checks the quality and legal risks of generated content, and further adjusts the content based on user sentiment, providing an environment where users can use it with peace of mind.

[1496] Example 2

[1497] 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."

[1498] The content generated by modern generative AI models may contain inappropriate content or risk copyright infringement. Furthermore, if the generated content does not match the user's emotional state, it can negatively impact the user experience. To address this issue, it is necessary to ensure the appropriateness of the content and take the user's emotional state into consideration.

[1499] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking the generated content if it contains inappropriate content, means for notifying the user about the generated content, means for recognizing the user's emotion, means for adjusting the content of the generated content based on the user's emotion, and means for suspending provision of the generated content based on the result of the user's emotion recognition. This makes it possible to provide content that matches the user's emotional state while ensuring the appropriateness of the generated content.

[1500] "Generated content" refers to information such as text, images, and video that is automatically generated by a generative AI model based on a user request.

[1501] "Filtering measures" are technologies or algorithms used to analyze generated content and determine whether it contains inappropriate content or is at risk of copyright infringement.

[1502] "Copyright infringement risk assessment means" means any technology or algorithm that compares generated content with existing databases or copyrighted material to assess the risk of copyright infringement.

[1503] "Blocking measures" are technologies or methods that remove generated content before it is provided to users if the content contains inappropriate content or poses a risk of copyright infringement.

[1504] A "notification mechanism" is a mechanism for informing users of the outcome of generated content, whether it is appropriate or blocked for being inappropriate.

[1505] "Emotion recognition means" refers to technology or algorithms that analyze a user's facial expressions, voice, and input data, and recognize the user's emotions in real time based on the results.

[1506] The "emotion-based content adjustment means" refers to a technology or algorithm for changing the content of generated content to match the emotional state of the user based on the result of the user's emotion recognition.

[1507] The "provision interruption means" is a technique or method for interrupting the provision of generated content midway when the user's emotion recognition result satisfies certain conditions.

[1508] MODE FOR CARRYING OUT THE INVENTION

[1509] This invention is a system that reduces the risk of inappropriate content and copyright infringement in content generated using a generative AI model, and combines it with an emotion engine that recognizes user emotions to provide more appropriate content. This system aims to further improve the quality of generated content and enhance the user experience by taking user emotions into consideration.

[1510] Hardware and software used

[1511] This system mainly uses the following hardware and software:

[1512] Server: A high-performance server for data analysis and processing. Specific examples include server instances on AWS or Google Cloud Platform.

[1513] Terminal: A device operated by a user. Examples include PCs, smartphones, tablets, etc.

[1514] Emotion engine: Software for analyzing a user's facial expressions and voice data. Examples include OpenCV and Microsoft Azure Emotion API.

[1515] Generative AI model: An artificial intelligence model for content generation. An example of this is OpenAI GPT-3.

[1516] Natural Language Processing (NLP) software, used to analyze generated content and check for inappropriate content and copyright infringement risks. Examples include spaCy and NLTK.

[1517] Specific operation of the system

[1518] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expressions and voice data through the connected camera and microphone. The collected data is analyzed in real time to recognize the user's emotions.

[1519] The user fills out a content generation request form for the AI ​​on their device screen and clicks the "Submit" button. The request includes the type of content they want to generate and guidelines.

[1520] The server receives content generation requests sent by users, formats them as necessary, and forwards them to the generative AI model, which generates content based on the requests and meets the specified conditions.

[1521] The server receives the content generated by the generative AI model, analyzes it using NLP software, and blocks it if it contains inappropriate content or is at risk of copyright infringement.

[1522] The emotion engine analyzes the user's emotional state and determines whether the content is appropriate for the user's state. If necessary, it adjusts the generated content. For example, if the user is feeling stressed, it changes the content to something more relaxing.

[1523] Finally, the server checks the generated content for appropriateness and provides it to the user. If the content is inappropriate, the server notifies the user of the result.

[1524] Specific examples

[1525] As an example, consider the case where a user submits a content generation request using the following prompt:

[1526] Example prompt sentence:

[1527] "Generate fun adventure stories for children, but please do not include dangerous or violent scenes."

[1528] Based on this request, a generative AI model generates an adventure story. The server analyzes the content to ensure it does not contain any dangerous scenes. Furthermore, an emotion engine analyzes the user's emotional state and adjusts the content as needed. The optimal content is then delivered to the user.

[1529] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1530] Step 1:

[1531] The server starts the emotion engine when the system starts up. The emotion engine collects the user's facial expression and voice data through the connected camera and microphone and analyzes it in real time. The input data is the user's facial expression and voice data, and the output is the analysis result that indicates the user's emotional state. Specifically, it calls the emotion engine's API and analyzes the collected data.

[1532] Step 2:

[1533] The user fills out a content generation request form for the generation AI on their device screen and clicks the "Submit" button. The input data is form data that includes the type of content they want to generate and guidelines, and the output data is the request data that is sent to the server. Specifically, the user fills out the form on the browser screen and sends a POST request.

[1534] Step 3:

[1535] The server receives a content generation request sent by the user. The received data is the user's request data, and the output is formatted request data to be sent to the generative AI model. The server formats the received request content as needed and sends it to the generative AI model. Specifically, it converts the received data into an internal format and sends it to the generative AI's API.

[1536] Step 4:

[1537] The generation AI generates content that meets the specified conditions based on the request. The input is the request data transferred from the server, and the output is the generated content data. Specifically, the generation AI runs a natural language generation model (e.g., GPT-3) based on the prompt sentence and outputs the generated results as text data.

[1538] Step 5:

[1539] The server receives the content generated by the generation AI. The input is the content data from the generation AI, and the output is the data to be analyzed for filtering and moderation. Specifically, the server receives the HTTP response and stores the generated content data in an internal database.

[1540] Step 6:

[1541] The server automatically filters and moderates the content it receives. The input is the content data to be analyzed, and the output is the filtering and moderation results. Specifically, it uses an NLP library (e.g., spaCy or NLTK) to analyze the text and check for inappropriate content and copyright infringement risks.

[1542] Step 7:

[1543] Based on the results of the emotion engine, the server determines whether the content is appropriate for the user's emotional state and adjusts the content as necessary. The input is the emotion recognition result and the content data to be analyzed, and the output is the adjusted content data. Specifically, the emotion recognition result is obtained from the API, and the content is regenerated or replaced with a template as necessary.

[1544] Step 8:

[1545] If the server determines that the generated content is inappropriate or poses a risk of copyright infringement, it blocks the content. The input is the filtering and moderation results, and the output is a list of blocked content. The specific operation is to add the content to the block list so that it is not provided to users.

[1546] Step 9:

[1547] The server notifies the user whether the generated content is appropriate or blocked because it is inappropriate. The input is the user's notification destination information and notification content data, and the output is the notification sending result. Specifically, the server notifies the user of the result via email or an in-app notification system.

[1548] (Application example 2)

[1549] 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."

[1550] Content generated by conventional generative AI models is not optimized based on user sentiment, and therefore does not provide a sufficiently good user experience. Furthermore, content may contain inappropriate content or risk copyright infringement, making it difficult to provide safe and appropriate content. This has led to problems such as reduced user satisfaction.

[1551] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1552] In this invention, the server includes means for automatically filtering generated content, means for evaluating the risk of copyright infringement of the generated content, means for blocking generated content if it contains inappropriate content, means for notifying users about the generated content, means for analyzing emotions of users, and means for adjusting the content of the generated content based on the emotions of users, thereby enabling the provision of appropriate content that takes into consideration the emotions of users.

[1553] "Means for automatically filtering generated content" refers to technology that automatically analyzes generated content and detects and removes inappropriate content or harmful information.

[1554] The "means for assessing the risk of copyright infringement of generated content" is a technology for assessing whether generated content uses existing copyrighted works without permission and determining the risk of copyright infringement.

[1555] "Means for blocking generated content if it contains inappropriate content" refers to technology that removes or stops the display of generated content if it contains inappropriate language or harmful information.

[1556] The "means for notifying users of generated content" is a technique for notifying users whether generated content is in an appropriate state when provided to the users.

[1557] "Emotion analysis means for recognizing user emotions" is a technology that analyzes the user's facial expressions, tone of voice, input data, etc. to recognize the user's emotional state in real time.

[1558] The "means for adjusting the content of generated content based on the user's emotions" is a technology for optimizing the content of generated content in accordance with the recognized user's emotions.

[1559] System Overview

[1560] The system for implementing this invention integrates multiple key functions, analyzes user emotions in real time, and appropriately filters and adjusts the generated content to provide high-quality content.

[1561] Program processing

[1562] The program of this system includes the following elements:

[1563] 1. Starting and analyzing the emotion recognition engine

[1564] The server collects data in real time from the camera and microphone of the user's device (smartphone, tablet, etc.), which allows it to analyze the user's facial expressions and voice and recognize their emotions.

[1565] The software used is a general image analysis API (e.g., image analysis cloud service) for image analysis, and an audio tone analysis tool (e.g., audio tone analysis service) for audio analysis.

[1566] 2. Submitting a content generation request

[1567] The user sends a request to the AI ​​via their device, including the type of content they want and keywords, and the request includes a specific prompt.

[1568] 3. Content generation using generative AI models

[1569] The server generates content based on specified conditions based on a generative AI model (e.g., a generative AI service).

[1570] The generated content is diverse, including text, images, and audio.

[1571] Specific examples

[1572] For example, if a user wants to send a request to create relaxing music, they might use a prompt like this:

[1573] Example prompt: "Generate relaxing music containing nature sounds that users would like to listen to while in a relaxed state."

[1574] Based on this request, the server issues instructions to the generation AI to generate the requested content.

[1575] 4. Filtering and Moderation

[1576] The server receives the generated content and automatically checks it for inappropriate content and risks of copyright infringement.

[1577] It utilizes natural language processing (NLP) to analyze text and uses machine learning libraries such as TensorFlow and PyTorch.

[1578] 5. Emotion-based content adjustment

[1579] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, the server regenerates content to provide soothing sensations without overstimulating them.

[1580] The parameters of the AI ​​model are adjusted based on the results of sentiment analysis to generate new content.

[1581] 6. Provision of Final Content and Notification

[1582] The server provides the tailored content to the user and notifies them if the content is appropriate.

[1583] If the user is in a relaxed state, content with a high relaxation effect is provided, improving the user experience.

[1584] As described above, the present invention takes into consideration the feelings of users and makes it possible to provide optimal content while reducing the risk of inappropriate content and copyright infringement.

[1585] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1586] Step 1:

[1587] The server collects data in real time from the camera and microphone of the device (smartphone, tablet, etc.). This data is used to obtain the user's facial expressions and voice data, and preprocesses it to recognize emotions. The collected data is analyzed using an image analysis cloud service and a voice tone analysis service, and the user's emotional state (e.g., "relaxed") is output.

[1588] Step 2:

[1589] The user inputs and sends a request through their device, including the type of content they want and keywords. The generated prompt is something like "Generate relaxing music." The server receives this request and prepares the prompt as data to send appropriate instructions to the generative AI model.

[1590] Step 3:

[1591] The server sends a prompt to the generative AI model, instructing it to generate content based on the specified conditions. The generative AI model generates content such as text, images, and audio based on the request. This generated content is then sent back to the server.

[1592] Step 4:

[1593] The server receives the generated content and automatically filters it for inappropriate content and copyright infringement risks. It uses natural language processing (NLP) to analyze the text and machine learning libraries (e.g., TensorFlow and PyTorch) to assess the risk. Based on this assessment, it blocks the content if it contains inappropriate content.

[1594] Step 5:

[1595] The server adjusts the generated content based on the results of the emotion analysis. For example, if the user is in a relaxed state, it checks whether the generated content is overly stimulating and regenerates it if necessary. It also adjusts the parameters of the AI ​​model to optimize the newly generated content.

[1596] Step 6:

[1597] The server provides the final adjusted content to the user and notifies them whether the content is appropriate. As the content is provided, the emotion engine analyzes it again in light of the user's emotions and provides appropriate feedback. If the user is relaxed, relaxation content is provided to maintain that state.

[1598] As described above, this system specifically realizes a mechanism for analyzing user emotions in real time and generating and providing optimal content based on that.

[1599] 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.

[1600] 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.

[1601] 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 robot 414.

[1602] 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.

[1603] 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.

[1604] 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.

[1605] 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).

[1606] 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.

[1607] 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."

[1608] 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.

[1609] 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).

[1610] 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.

[1611] 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.

[1612] 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.

[1613] 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.

[1614] 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.

[1615] 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.

[1616] 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.

[1617] 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.

[1618] 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.

[1619] 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.

[1620] The following is further disclosed regarding the above embodiment.

[1621] (Claim 1)

[1622] a means for automatically filtering generated content;

[1623] a means of assessing the risk of copyright infringement of generated content;

[1624] A means to block generated content if it contains inappropriate material; and

[1625] a means for notifying a user of the generated content;

[1626] A system including:

[1627] (Claim 2)

[1628] 10. The system of claim 1, wherein the analysis of the generated content is performed using natural language processing.

[1629] (Claim 3)

[1630] 10. The system of claim 1, wherein the system matches the similarity of the generated content with an existing database.

[1631] "Example 1"

[1632] (Claim 1)

[1633] A means for a user to use a terminal to send a content generation request to the generation AI;

[1634] A means for the server to receive requests from users and forward them to the generating AI;

[1635] A means for the generative AI to generate content based on requests;

[1636] A means for the server to receive the generated content from the generation AI;

[1637] means for automatically filtering and moderating server-generated content;

[1638] A means to block server-generated content if it is deemed inappropriate or poses a risk of copyright infringement; and

[1639] a means for the server to notify the user of the result of whether the generated content is appropriate or blocked;

[1640] A system including:

[1641] (Claim 2)

[1642] 10. The system of claim 1, wherein the analysis of the generated content is performed using natural language processing.

[1643] (Claim 3)

[1644] 10. The system of claim 1, wherein the system matches the similarity of the generated content with an existing database.

[1645] "Application Example 1"

[1646] (Claim 1)

[1647] a means for automatically filtering generated content;

[1648] a means of assessing the risk of copyright infringement of generated content;

[1649] A means to block generated content if it contains inappropriate material; and

[1650] a means for notifying a user of the generated content;

[1651] A means of checking generated content before distribution;

[1652] means for analyzing audio and subtitles of generated content using natural language processing;

[1653] An evaluation method for generated content based on guidelines specific to educational content;

[1654] A system including:

[1655] (Claim 2)

[1656] 10. The system of claim 1, wherein the analysis of the generated content is performed using natural language processing.

[1657] (Claim 3)

[1658] 10. The system of claim 1, wherein the system matches the similarity of the generated content with an existing database.

[1659] "Example 2: Combining Emotion Engines"

[1660] (Claim 1)

[1661] a means for automatically filtering generated content;

[1662] a means of assessing the risk of copyright infringement of generated content;

[1663] A means to block generated content if it contains inappropriate material; and

[1664] a means for notifying a user of the generated content;

[1665] means for recognizing a user's emotion;

[1666] means for adjusting the content of the generated content based on the user's emotions;

[1667] a means for suspending provision of generated content based on a result of the user's emotion recognition;

[1668] A system including:

[1669] (Claim 2)

[1670] 10. The system of claim 1, wherein the analysis of the generated content is performed using natural language processing.

[1671] (Claim 3)

[1672] 10. The system of claim 1, wherein the system matches the similarity of the generated content with an existing database.

[1673] "Application example 2 when combining emotion engines"

[1674] (Claim 1)

[1675] a means for automatically filtering generated content;

[1676] a means of assessing the risk of copyright infringement of generated content;

[1677] A means to block generated content if it contains inappropriate material; and

[1678] a means for notifying users of generated content;

[1679] emotion analysis means for recognizing the emotions of a user;

[1680] A means for adjusting the content of the generated content based on the user's emotions;

[1681] A system including:

[1682] (Claim 2)

[1683] 10. The system of claim 1, wherein the analysis of the generated content is performed using natural language processing.

[1684] (Claim 3)

[1685] 10. The system of claim 1, wherein the system matches the similarity of the generated content with an existing database. [Explanation of symbols]

[1686] 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 means for automatically filtering generated content; a means of assessing the risk of copyright infringement of generated content; A means to block generated content if it contains inappropriate material; and a means for notifying a user of the generated content; A system including:

2. The system of claim 1 , wherein the analysis of the generated content is performed using natural language processing.

3. The system of claim 1 , wherein the system matches the similarity of the generated content with an existing database.

Images