AI-generated content management system, method, and program

Abstract

This enhances the traceability of generated content, prevents the spread of fraudulent content such as deepfakes, and improves the reliability of AI-generated content. [Solution] The system includes: an audio watermark authentication data insertion processing unit 36d that inserts unique identification information for the AI-generated content as an audio watermark into the acoustic signals contained in the AI-generated content when generating AI-generated content; an authentication data insertion processing unit 36d that acquires tracking data including information related to the generation of the AI-generated content; a blockchain that stores at least a portion of the authentication information that identifies the audio watermark in association with traceability data D1; and a tracking information inquiry unit 36b that, when playing the AI-generated content C1, extracts the audio watermark inserted into the AI-generated content related to the playback and queries tracking data from the blockchain based on the extracted audio watermark.

Description

Technical Field

[0001] The present invention relates to a technology for managing content generated based on information learned by artificial intelligence, and an AI-generated content management system, method, and program capable of enhancing the traceability of AI-generated content, preventing the spread of illegal content, and improving the reliability of AI-generated content.

Background Art

[0002] In recent years, AI technology has been evolving rapidly, and its progress has created new possibilities in various fields. In particular, by using technologies such as deep learning and natural language processing, it has become possible to generate extremely sophisticated videos and voices. By applying this technology, the costs and time related to conventional media production can be significantly reduced, and high-quality content can be provided in a shorter period. For example, cases where an AI announcer automatically reads news or advertisements, or a virtual influencer is active on social media, adding new value to a company's marketing, are increasing. Thus, the development of AI technology has advanced in various fields, from entertainment, advertising, education to daily life.

[0003] On the other hand, the social impact of deepfakes is also increasing. Deepfake technology can create realistic false information, fabricated videos, and voices, which may deceive people. In particular, the risk of the spread of political false information and fake videos of famous people on social media is increasing, raising concerns about a decline in reliability and social chaos, and cases causing significant economic losses have also been reported.

[0004] As a conventional technology for detecting such deepfakes, a technology for embedding a watermark in an audio signal that constitutes digital content such as a video or audio is known. For example, Patent Document 1 discloses a technology related to a watermark for preventing and detecting forgery of an audio signal.

Prior Art Documents

[0005] [Patent Document 1] Japanese Patent Publication No. 2004-310117 [Overview of the project] [Problems that the invention aims to solve]

[0006] However, while the aforementioned conventional technologies provide a mechanism for verifying the authenticity of audio and video signals by embedding watermarks, they have the problem of not adequately considering the traceability of the generated digital content.

[0007] Specifically, even if watermarking technology can identify that content is AI-generated, it is not possible to track who generated or edited the content, when, or for what purpose. This makes it difficult to identify the source, and there is a problem in that it is not possible to confirm whether the AI-generated content was legitimately distributed by a person or organization, or whether it is malicious content such as deepfake.

[0008] The lack of traceability in such conventional technologies increases the risk of serious social and economic damage, especially when misused for political misinformation or corporate fraud. On the other hand, the aforementioned conventional watermarking technologies do not ensure traceability, making it difficult to verify the authenticity of content and potentially diluting the credibility of legitimate AI-generated content.

[0009] Therefore, the present invention aims to solve the above-mentioned problems and to provide an AI-generated content management system, method, and program that can prevent the spread of fraudulent content such as deepfakes, improve the reliability of AI-generated content, and prevent social disruption and economic losses by enhancing the traceability of generated content, thereby enabling the recording and tracking of who generated the content, when, and for what purpose. [Means for solving the problem]

[0010] To solve the above problems, the present invention provides a system for managing content generated based on information learned by artificial intelligence as AI-generated content, When generating the aforementioned AI-generated content, an audio watermark insertion unit inserts unique identification information for the AI-generated content as an audio watermark into the audio signal contained in the AI-generated content. A tracking data acquisition unit that acquires tracking data including information related to the generation of the aforementioned AI-generated content, A data storage means that stores at least a portion of the authentication information that identifies the aforementioned voice watermark in association with the aforementioned tracking data, When playing the AI-generated content, the inquiry unit extracts the audio watermark inserted within the AI-generated content being played, and queries the tracking data from the data storage means based on the extracted audio watermark. It is equipped with.

[0011] Furthermore, the present invention relates to a method for managing content generated based on information learned by artificial intelligence as AI-generated content, (1) When generating the AI-generated content, the audio watermark insertion unit inserts an identification information unique to the AI-generated content as an audio watermark into the acoustic signal contained in the AI-generated content; (2) A data storage step in which tracking data including information relating to the generation of the AI-generated content is acquired, and at least a portion of the authentication information identifying the voice watermark is stored in association with the tracking data in the data storage means, (3) When playing the AI-generated content, the inquiry unit extracts the audio watermark inserted in the AI-generated content related to the playback, and queries the tracking data from the data storage means based on the extracted audio watermark in an inquiry step. Includes.

[0012] In the above invention, it is preferable that the tracking data acquisition unit acquires identifiers and variables defined for each item as tracking data related to the generation of the AI-generated content. Furthermore, in the above invention, it is preferable that the audio watermark is inserted as an inaudible sound within the acoustic signal included in the AI-generated content, and the inquiry unit acquires the identification information by extracting the inaudible sound. Moreover, it is preferable that the data storage means comprises a plurality of nodes that encrypt and store at least a portion of the tracking data, the nodes aggregate and block at least a portion of the tracking data at a predetermined timing, link these blocks to existing blocks to form a blockchain, and share the blockchain among the plurality of nodes and store it as a distributed ledger.

[0013] Furthermore, the systems and methods according to the present invention described above can be realized by executing a program of the present invention written in a predetermined language on a computer. That is, by installing the program of the present invention on the IC chip or memory device of a mobile terminal device, smartphone, wearable device, mobile PC or other information processing terminal, or a general-purpose computer such as a personal computer or server computer, and executing it on the CPU, a system having the above-described functions can be constructed, and the methods according to the present invention can be implemented.

[0014] The program of the present invention can be distributed, for example, via a communication line, and can also be transferred as a packaged application that runs on a standalone computer by recording it on a computer-readable recording medium. Specifically, this recording medium can be various recording media, such as magnetic recording media, optical discs, and RAM cards. Furthermore, by using a computer-readable recording medium on which this program is recorded, the above-described system and method can be easily implemented using a general-purpose computer or a dedicated computer, and the program can be easily stored, transported, and installed. [Effects of the Invention]

[0015] As described above, these inventions assign a unique audio watermark to AI-generated content, and store tracking data containing information about the content's creation (who generated it, when, and for what purpose, etc.) and authentication information identifying the audio watermark in a data storage means. This system makes it possible to verify the origin and generation path of AI-generated content afterward, thereby ensuring the authenticity and transparency of the source of the AI-generated content. As a result, it is expected that fraudulent content such as deepfakes will be eliminated from the market, and highly reliable content generation technologies will become widespread.

[0016] In particular, since this invention simultaneously inserts a unique watermark (audio fingerprint) when content is generated, unlike conventional techniques that insert watermarks afterward, the origin and history of the generated audio can be traced more reliably. As a result, even if the generated audio is misused by a third party, the source can be accurately identified and responsibility can be clearly determined.

[0017] In addition, according to the present invention, by means of the tracking data simultaneously provided during content generation, all the histories of the generated content can be tracked in a database. Enterprises and users utilizing AI generation technology can confirm how the generated voice has been used and in which applications or solutions it has been utilized, making it possible to prove the legitimacy of the voice. Moreover, by providing generation and query in a set, the reliability of voice generation technology can be significantly improved.

[0018] The present invention can establish traceability in AI-generated content, reduce the risk of fake voices circulating in the market, and provide an environment in which enterprises and individuals can safely utilize AI voice generation technology. The reliability in the voice generation market is enhanced, enabling large enterprises and organizations to utilize AI-generated content with confidence.

[0019] Note that AI-generated content is classified into three types: "real voice", "voice generated by AI to be exactly like the real one", and "generated voice by pure AI". Regardless of this classification, it is possible to accurately distinguish whether the content is generated by AI or whether there is approval from the producer himself / herself.

Brief Description of the Drawings

[0020] [Figure 1] It is a conceptual diagram showing an overview of AI traceability in a content generation service according to an embodiment. [Figure 2] It is a block diagram showing the overall configuration of an AI-generated content management system according to an embodiment. [Figure 3] It is an explanatory diagram showing an overview of content tracking in an AI-generated content management system according to an embodiment. [Figure 4] It is an explanatory diagram showing an overview of content tracking in other AI content generation services. [Figure 5] It is a block diagram showing the internal configuration of an AI content management server according to an embodiment. [Figure 6] It is a block diagram showing the internal configuration of a guarantee system according to an embodiment. [Figure 7] This is a block diagram showing the internal configuration of a smartphone according to an embodiment. [Figure 8] This is a sequence diagram showing the operation of the AI-generated content management service according to the embodiment. [Figure 9] This is an explanatory diagram illustrating the process of registering and retrieving AI-generated content using blockchain technology in an embodiment. [Figure 10] This is an explanatory diagram showing the data structure of traceability data D1 according to the embodiment. [Modes for carrying out the invention]

[0021] Embodiments of the AI-generated content management method, management system, and management program according to the present invention will be described in detail below with reference to the attached drawings. Note that the embodiments shown below are illustrative examples of devices, etc., for realizing the technical concept of this invention, and the technical concept of this invention does not limit the materials, shapes, structures, arrangements, etc., of each component to those described below. Various modifications can be made to the technical concept of this invention within the scope of the claims.

[0022] (Overview of AI traceability in AI generation services) In the AI ​​generation service according to this embodiment, as shown in Figure 1, when the generation system AI 6 automatically generates content including audio signals such as video and audio (F1), the generation system AI 6 simultaneously adds a unique audio watermark to the content (F2). Specifically, at the stage of AI content generation, a uniquely recognized ID specific to that content is inserted as an inaudible audio watermark into the audio signals included in the content generated by the generation system AI 6.

[0023] This uniquely identified ID is used as a traceability code, and this traceability code, when combined with other tracking data items, constitutes traceability data. More specifically, traceability data is formed by concatenating the traceability code with variables attached, and this traceability code includes information about the creation of the content (who created it, when, and for what purpose, etc.), as illustrated in Figure 10. Such traceability data and authentication information that identifies the audio watermark (uniquely identified ID and recognition code at the time of NFT registration) are linked to each other and registered (stored) in an AI-generated content database or blockchain, which are data storage means.

[0024] In this embodiment, the generated AI-generated content, in whole or in part (for example, only the audio portion), is registered as original data in a blockchain-based NFT (Non-Fungible Token) system for authenticity. The unique recognition code assigned to the content upon registration with the NFT, along with the traceability code incorporating the unique recognition ID, are used as authentication information to identify the AI-generated content and the audio watermark. Here, the recognition code from NFT registration and the unique recognition ID of the audio watermark are defined as the "AI Trace Code," which is the query code within the traceability code, thereby linking the traceability code (tracking data) with the authentication information that identifies the audio watermark.

[0025] Furthermore, in this embodiment, in addition to watermarking with inaudible sound, the voiceprint of the person (or character) speaking in the content is also used to identify the content. The traceability database 35b and AI-generated content database 35c shown in Figure 5 have registered voiceprints of famous people and famous characters (including AI-generated ones), and it is possible to check whether the voice is that of the person by matching these voiceprints.

[0026] As shown in Figure 3, by querying tracking data (traceability code) via the voice watermark and voiceprint, it becomes possible to track the history of AI-generated content in detail using information related to the creation of this content. The history related to such voice watermarks and the voiceprint of the audio in the content are unique traces / voiceprints of the generated content, and by analyzing them and querying the database, it is possible to uniquely identify which AI created the audio and for whom. For example, as shown in Figure 3, if user U1 creates video content C1 of a news program featuring a famous female announcer with audio using the generation system AI 6, and the voiceprint matches, it can be determined that the female announcer herself or someone who looks very similar (including AI-generated) is reporting the news, but it is not actually her speaking. In this case, information such as whether or not her approval was obtained is written in traceability data D1, linked to the voice watermark, and registered in the traceability database 35b and the blockchain.

[0027] When AI-generated content with such voice watermarks is distributed and used (F3 in Figure 1), it becomes possible to track which applications and solutions the content is being used in based on the voice watermark and traceability data (F4). After the AI-generated content is released to the market, if a third party wishes to verify the authenticity and creation history of that AI-generated content, they can access the AI-generated content management server 3 in the tracking and verification phase to check the creator and creation status from the voice watermark and voiceprint. This proves whether the AI-generated content is genuine, generated by AI, or is pure AI voice.

[0028] In the example shown in Figure 3, when video content C1 of a news program is distributed and viewed by user U2, the traceability inquiry function of the browser software extracts an audio watermark from the sound signal in the video, reads the associated traceability data D1 from the traceability database 35b or blockchain, and compares the retrieved traceability data D1 with the content of the audio watermark to prove the authenticity of the content. If they match, the authenticity is proven. If the audio watermark has been deleted or tampered with, or if the content itself has been altered, the audio watermark and traceability data D1 will not match, making proof impossible, and a warning will be issued to user U2.

[0029] On the other hand, as shown in Figure 4, if an external user Ux who does not use this system creates similar video content Cx using AI, the insertion of audio watermarks and registration of traceability data on the blockchain, as performed by this system, will not occur. Therefore, when user U2 attempts to view this video content Cx, even immediately after distribution, the traceability data to be queried does not exist on the blockchain, making verification impossible, and a warning will be issued during playback. In this case, even if the video content Cx is a copy, or if audio watermarks are later illegally inserted or the content itself is altered, the fact remains that the traceability data to be queried does not exist on the blockchain. In either case, verification will be impossible, and a warning will be issued during playback.

[0030] Ultimately, this information is fed back into the generative AI's further generation process (F1), and the generation-verification cycle is repeated, ensuring transparency and reliability of content generation by the generative AI. In this system, what is important is not whether the people in the video content are real or fake, but rather the transparency of the generation process and whether the audio has been approved. By providing a mechanism that allows anyone to trace (verify) this process, transparency and reliability of content generation by the generative AI are guaranteed. Content that cannot be verified in this way, regardless of its authenticity, will reduce the credibility of the distributor if it is used, and it is expected that fraudulent content such as deepfakes will be eliminated from the market.

[0031] (Overall configuration of the AI-generated content management system) In this embodiment, the AI ​​content generation service described above is provided by an AI content generation management system as shown in Figure 2. As shown in the figure, in this service, user U1 generates content for the AI ​​generation system 6 via the AI ​​content generation management server 3 through the user's smartphone 5. The generated AI content is managed and operated by the AI ​​content generation management server 3.

[0032] Here, AI-generated content refers to content generated using generative artificial intelligence (AI), specifically content such as audio, images, videos, and text that is automatically generated based on data learned by the AI. Examples include videos in which real people, people who have lived in the past, or fictional people speak and move. In the case of videos of real people, this includes content in which the person actually spoke, or has approved or denied the AI ​​generation of their voice, movements, appearance / clothing, and speech, and can take the following forms. (1) Audio that the voice owner has acknowledged as being their own voice (their own voice) (2) An AI-generated voice that is very similar to (1) (3) A voice generated by pure AI that belongs to no one (a fictional voice generated by AI)

[0033] The AI-generated content management system 1 according to this embodiment is configured such that an AI-generated content management server 3, which generates and manages AI-generated content, and a smartphone 5, which is an information processing terminal used by each user, are interconnected via a communication network 2. In this embodiment, the smartphone 5 is described as an example of an information processing terminal. Furthermore, a blockchain platform 4, which is a guarantee system that provides a blockchain interface service to guarantee AI-generated content management, is installed on the communication network 2, and it is possible to connect to a distributed ledger system such as a blockchain through this blockchain platform 4.

[0034] Communication Network 2 is an IP network such as the Internet that uses the TCP / IP communication protocol, and is a distributed communication network constructed by interconnecting various communication lines (public lines such as telephone lines and fiber optic lines, dedicated lines, fourth-generation (4G) communication methods such as LTE, and fifth-generation (5G) and later communication methods, as well as wireless communication networks such as Wi-Fi® and Bluetooth®). This IP network also includes LANs such as intranets (corporate networks) and home networks using 10BASE-T and 100BASE-TX.

[0035] The AI-generated content management server 3 is a server device that provides the generation, management, and use of AI-generated content as a service. In this embodiment, it has the function of generating, saving, and distributing AI-generated content data D2 via a smartphone 5, and can be realized, for example, by a single server device or a group of multiple server devices. The AI-generated content management server 3 is configured to virtually construct multiple functional modules on the CPU, with each functional module working together to perform various processes. Furthermore, the AI-generated content management server 3 can send and receive data via the communication network 2 using its communication function, and can also display websites and provide services via browser software using its web server function.

[0036] On the other hand, the smartphone 5 is a portable information processing terminal device that uses wireless communication. It communicates wirelessly with relay points such as wireless base stations 22, and can receive communication services such as voice calls and data communication while on the move. Examples of this communication method include 3G (3rd Generation), LTE (Long Term Evolution), 4G, FDMA, TDMA, CDMA, W-CDMA, and PHS (Personal Handyphone System). Furthermore, the smartphone 5 is equipped with various functions such as a digital camera, application software execution function, and location information acquisition function using GPS (Global Positioning System), and also includes mobile computers such as tablet PCs.

[0037] The location information acquisition function for the smartphone 5 is a function that acquires and records location information indicating the location of the device. This location information acquisition function includes methods such as detecting the device's position using signals from satellites, like GPS, or detecting the position using radio wave strength from cellular base stations 22 or Wi-Fi access points.

[0038] The smartphone 5 is equipped with a display unit 53a, such as a liquid crystal display, for displaying information, and an operation device such as an operation button for the user to perform input operations. This operation device includes a touch panel, which is superimposed on the liquid crystal display and acquires operation signals such as touch operations that specify coordinate positions on the liquid crystal display. Specifically, this touch panel is an input device that receives operation signals by pressure, electrostatic detection, etc., from touch operations using the user's fingertip or a pen, and is configured by superimposing a liquid crystal display that displays graphics and a touch sensor that receives operation signals corresponding to the coordinate positions of the graphics displayed on the liquid crystal display.

[0039] (Internal structure of each device) Next, the internal structure of each device constituting the AI-generated content management system 1 described above will be explained. Figure 5 shows the AI-generated content management server 3 according to this embodiment, and Figure 7 shows the smartphone 5 according to this embodiment. In this explanation, the term "module" refers to a functional unit composed of hardware such as devices or equipment, software with the same functionality, or a combination thereof, for achieving a predetermined operation.

[0040] (1) AI-generated content management server 3 First, let's explain the internal configuration of the AI-generated content management server 3. The AI-generated content management server 3 is a server device located on the communication network 2, and it is capable of sending and receiving data with other information terminals such as smartphones 5, blockchain interfaces, and various other servers via the communication network 2.

[0041] Specifically, the AI-generated content management server 3 includes a communication interface 31 that performs data communication via a communication network 2, an authentication unit 33 that authenticates the authority of users and user terminals, a location information management unit 32 that collects and manages location information of each user terminal, a traceability management unit 36 ​​that performs control such as generation, storage and management of AI-generated content data, a content distribution unit 34 that distributes information to each user, and various database groups 35a to 35c.

[0042] The database group includes a user database 35a for storing information about users, a traceability database 35b, and an AI-generated content database 35c as a data storage means for storing generated AI-generated content. Each of these databases may be a single database, or it may be divided into multiple databases and relationships may be established between them to create a relational database that links the data together.

[0043] The information stored in the user database 35a includes information that enables individual recognition of users who generate AI-generated content. This may include personal information such as names, user attribute information such as age, gender, and occupation, purpose of use and content usage status, generation date and time, type of generated content, generation history information of the models and algorithms used, and feedback provided on the generated content. The information stored in the user database 35a also includes authentication information linked to identifiers (user ID, terminal ID) that identify the user or the information processing terminal used by the user, and passwords, etc. This also includes personal information of the user linked to the user ID and the model of the terminal device. Furthermore, the user database 35a also stores authentication history (access history) for each user or user terminal, and information (including related information) about AI-generated content for each user through its relationship with the AI-generated content database 35c.

[0044] Furthermore, the user database 35a can also store image information of official identification documents such as driver's licenses, passports, and My Number cards for the purpose of detailed user authentication and identity verification. The user database 35a can also store image data and text data extracted from these identification documents, and information such as the facial photograph, name, address, issue date, and expiration date of the driver's license, the facial photograph, name, nationality, issue date, and expiration date of the passport, and the facial photograph, name, address, issue date, expiration date, and My Number of the My Number card may be stored as user information. Since this personal information is highly confidential, it is encrypted and stored using strong encryption technology. Such image information and text data of identification documents are used to ensure security and meet legal requirements in service use, such as verifying the user's identity, age, and identity during transactions. For example, when buying, selling, or transferring AI-generated content, this information is referenced to verify the identities of both parties in the transaction and prevent fraud.

[0045] The traceability database 35b is a storage device that stores traceability data as tracking data containing information about the generation of AI-generated content, as shown in Figure 10. The traceability data is related to the AI ​​generation of content. Who • What is the purpose? ·when What kind of solution What kind of app? • How much (in seconds, in characters)? This is data that records how it was generated.

[0046] The traceability data shown in Figure 10 is as follows: • Timestamp (TIM) This indicates the time the AI-generated content was registered and the time of each generation, such as "2024 / 09 / 05 14:35:26". ·Generated by (MAK), generated by (ORG) This includes information that identifies the producer, such as the location, IP address, and name (optional) of the content registrant, as well as information that identifies the person (owner of the voice) from whom the audio in the content was recorded, such as the name of the organization to which the person belongs, as in "Former Prime Minister AB". • Personal ID code (1 or 2) This code indicates the approval status of the content, and can be a number or string, such as 1 if the person has approved it, and 2 otherwise. If the person has approved it, it includes not only audio and video of the person actually speaking, but also audio and video that sound exactly like the person, even if it is generated by an AI that the person has never actually spoken, as long as the person has acknowledged the content and given permission for it to be distributed under their name. ·Generation destination (SLU) This information, such as "AI-generated character solution," indicates the type of solution in which the AI-generated content is used. It includes information that identifies the service solution, such as whether it is a service aimed at generating realistic content based on the voice and video of real people, or a service aimed at creating virtual characters. ·Product (APP), production amount (GEN) This includes product information indicating which applications it is used in, such as an "AI tourist guide app," and information indicating the total amount of video and audio generated by the AI, such as "seconds or character count." AI Trace Code (TRC), Voice Generation AI Code (AIV(N)) This includes information such as a reference code for tracing the generated content, like "AB Office," and unique codes for each of the many generated pieces of content.

[0047] The AI-generated content database 35c described above is a data storage means for storing data of AI-generated content. The data stored in this AI-generated content database 35c includes content generated by users using the AI ​​generation system 6, stored as data files, as well as video and audio files uploaded by each user as source material.

[0048] The authentication unit 33 is a module that establishes a communication session with an information and communication terminal that wishes to use the content generation service via the communication interface 31, and performs authentication processing for each established communication session. This authentication process involves obtaining authentication information from the user's smartphone 5a, referring to the user database 35a to identify the user, and authenticating their authority. The authentication results from the authentication unit 33 (user ID, authentication time, session ID, etc.) are sent to the traceability management unit 36 ​​and stored in the user database 35a as authentication history.

[0049] The location information management unit 32 is a module that acquires location information obtained on the user's smartphone 5 and transmitted to the AI-generated content management server 3. The location information management unit 32 links the user and user terminal device identifiers (user ID, terminal ID, etc.) identified by the authentication process performed by the authentication unit 33 with their location information and stores them in the user database 35a as usage history.

[0050] The traceability management unit 36 ​​is a module that performs management and processing related to content generation, distribution, tracking, and inquiry. Specifically, the traceability management unit 36 ​​comprises a generation processing execution unit 36a, a tracking information inquiry unit 36b, and a guarantee system linkage unit 36c.

[0051] The generation processing execution unit 36a is a module that performs content generation processing based on the information learned by the artificial intelligence. The processing performed by the generation processing execution unit 36a during AI content generation is as follows: • Receiving a generation request This process receives a content generation request (generation request) from the user to the generative AI6. This request includes necessary generation parameters (types such as text, images, and audio, desired style and format), prompts, and more. • Data preprocessing This process involves converting user requests into a format suitable for processing by the AI ​​model, as well as pre-processing such as formatting input data, noise reduction, and extraction of necessary information. • Execution of AI model After the above preprocessing, the trained AI model generates content based on the preprocessed data. At this time, the authentication data insertion processing unit 36d inserts an audio watermark and registers traceability data (tracking data). • Post-processing of generated content Next, the generated AI content is checked to ensure it is appropriate and maintains the required quality, and adjustments such as filtering and formatting are made as needed. Afterward, the completed content is returned to the user. At this time, the processing status of the request (success, failure, etc.) is also provided via an HTTP response status code.

[0052] The authentication data insertion processing unit 36d, provided in the generation processing execution unit 36a, is an audio watermark insertion unit that, when generating AI-generated content, inserts identification information unique to the AI-generated content as an audio watermark into the acoustic signals contained in the AI-generated content. In inserting this audio watermark, the authentication data insertion processing unit 36d also functions as a tracking data acquisition unit that acquires traceability data, which is tracking data containing information related to the generation of the AI-generated content.

[0053] The tracking information inquiry unit 36b is a module that, when playing AI-generated content, extracts audio watermarks inserted within the AI-generated content being played, and queries tracking data from the blockchain and traceability database 35b, which are data storage means, based on the extracted audio watermarks. In this embodiment, the audio watermarks are inserted as inaudible sounds within the acoustic signals contained in the AI-generated content. The tracking information inquiry unit 36b obtains the audio watermark, which is identification information, by extracting the inaudible sounds within the AI-generated content, and queries the traceability data, which is tracking data, based on the obtained audio watermarks.

[0054] The assurance system integration unit 36c is a module that requests the blockchain platform 4 on the network to perform processing necessary for AI-generated content management, such as information related to AI-generated content, security management, and storage of transfer records, and collaborates with the blockchain platform 4 to carry out the processing. The traceability management unit 36 ​​manages the generation, tracking, inquiry, and certification of AI-generated content by coordinating with the blockchain platform 4 through this assurance system integration unit 36c. For example, it is possible to transfer ownership by adding a public address related to the acquirer of AI-generated content and changing the owner of various rights certified on the blockchain.

[0055] Blockchain platform 4 comprises multiple nodes that encrypt and store at least a portion of the tracking data of AI-generated content data. These nodes link voice data and related information contained in the AI-generated content data, aggregate at least a portion of that information at predetermined times, and block it. This block is then linked to existing blocks to form a blockchain, which is shared among multiple nodes and stored as a distributed ledger.

[0056] The content distribution unit 34 is a module that, in response to user access, presents a website related to the AI-generated content management service, provides each user with a unique My Page, and distributes AI-generated content and related information owned by each user individually. It also has a function to present an NFT trading platform for registering and trading AI-generated content as NFTs, and distributes the graphics and user interface of the NFT trading platform related to the generated AI-generated content to each user via the communication interface 31.

[0057] (2) Smartphone 5 Next, the internal configuration of smartphone 5 will be described. Smartphone 5a of user U1, who generates and distributes AI-generated content, has a generation application installed to realize AI traceability during content generation. On the other hand, smartphone 5b of user U2, who views the generated AI-generated content, has a viewing application installed to realize AI traceability during viewing. These applications may be dedicated applications that individually implement generation and viewing, applications that have both generation and viewing capabilities, or applications implemented as extensions of other browser applications. In this embodiment, we will illustrate a case where smartphone 5 is a device that can be used by both users U1 and U2, and has applications that have both generation and viewing capabilities installed.

[0058] As shown in Figure 7, the smartphone 5 is generally composed of a communication interface 51, an input interface 52, an output interface 53, an application execution unit 54, and a memory 55. The communication interface 51 is a communication device for making calls or data communications, and has the function of contactless communication via wireless or other means, as well as contact (wired) communication via cables, adapters, etc. The input interface 52 is a device for inputting user operations, such as a mouse, keyboard, operation buttons, or touch panel 52a. The output interface 53 is a device for outputting video and audio, such as a display or speaker. In particular, this output interface 53 includes a display unit 53a, such as an LCD display, and this display unit is superimposed on the touch panel 52a, which is the input interface.

[0059] Memory 55 is a storage device that stores the OS (Operating System), firmware, programs for various applications, and other data. This memory 55 stores user IDs to identify users, application data downloaded from the AI-generated content management server 3, and various data processed by the application execution unit 54. In particular, in this embodiment, memory 55 stores data acquired from the AI-generated content management server 3.

[0060] The application execution unit 54 is a module that executes applications such as general operating systems, game applications, and browser software, and is usually implemented by the CPU or the like. In this embodiment, an application equipped with both generation and viewing functions is installed, and a functional module that can be used in common by both users U1 and U2 is implemented. When the program of the AI-generated content management application according to the present invention is executed in this application execution unit 54, the AI ​​generation operation unit 546, the synchronization processing unit 543, the display control unit 545, and the location information acquisition unit 544 are virtually constructed.

[0061] The AI ​​generation operation unit 546 is a module that has functions for generating, downloading, and distributing AI-generated content on the smartphone 5. The processes operated by the AI ​​generation operation unit 546 when generating AI content are as follows: • Sending a generation request This process involves the user sending a content generation request (generation request) to the AI ​​model. For example, when creating a video of a real person speaking, the user sends necessary generation parameters (types such as text, images, and audio, desired style and format), such as prompts to identify the person's name, speech content, clothing, and facial expressions. During this generation operation, traceability data is also created, and necessary items are entered and modified. Upon sending this generation request, the trained AI model generates the content. At this time, an audio watermark is inserted, traceability data is registered, and the AI-generated content is registered as an NFT.

[0062] • Distribution of AI-generated content This module accepts user input to deliver AI-generated content and related information. Upon receiving this input, the content distribution unit 34 on the AI-generated content management server 3 delivers the AI-generated content. It also allows for operations such as registering and trading AI-generated content as NFTs.

[0063] The content viewing unit 540 is a module that has the function of viewing AI-generated content on the smartphone 5. Here, the content viewing function is provided in a dedicated application that is common to both the AI ​​generation operation and the content viewing function, but this content viewing function may also be a separate application or implemented as an extension of existing browser software. The content viewing unit 540 has a watermark data extraction unit 541, an authenticity inquiry unit 542, and a synchronization processing unit 543.

[0064] The watermark data extraction unit 541 is a module that extracts audio watermarks inserted into AI-generated content when playing the AI-generated content, and has the function of extracting audio watermarks inserted as inaudible sounds within the acoustic signals of the AI-generated content.

[0065] The authenticity inquiry unit 542 is a module that requests tracking data from the AI-generated content management server 3 based on the audio watermark extracted by the watermark data extraction unit 541. The authenticity inquiry unit 542 also has the function of outputting the results to the user via the display control unit 545.

[0066] The synchronization processing unit 543 is a module that synchronizes the AI-generated content management processing on the smartphone 5 with the AI-generated content management processing on the AI-generated content management server 3. Specifically, the AI-generated content management server 3 predicts possible event processing based on the positions of other users' characters and objects, generates the conditions for these events on the AI-generated content management server 3, sends these conditions to the smartphone 5, receives these conditions in the synchronization processing unit 543, and the actual event processing and the graphics processing for it are executed by the GUI control unit 545a on the smartphone 5 based on the conditions received from the AI-generated content management server 3. The results of the event processing executed by the GUI control unit 545a on the smartphone 5 (such as the outcome of battles and minigames, and scores) are notified to the GUI control unit 545a on the AI-generated content management server 3 via the synchronization processing unit 543 and reflected in the subsequent game progression processing.

[0067] The display control unit 545 is a module that performs specific screen display control based on the data generated by the display control unit 545. In this embodiment, the display control unit 545 has a GUI control unit 545a, which enables interaction with the user via the touch panel 52a and the display unit 53a, and performs appropriate screen transitions and information updates in response to user operations.

[0068] Furthermore, the display control unit 545 also has a function to generate display data for display on the display unit 53a. This display data includes graphic data, as well as data generated by combining image data, text data, video data, audio, and other data. For example, screens such as a list display of AI-generated content or a detailed display of specific AI-generated content data are generated as display data.

[0069] The GUI control unit 545a is a module responsible for displaying and controlling the graphical user interface (GUI) on the smartphone 5's display. This GUI control unit 545a displays various graphical elements such as buttons, menus, icons, and windows on the display to allow users to intuitively operate applications, and performs appropriate responses and actions in response to user input such as touch operations. The GUI control unit 545a operates in conjunction with the display control unit 545.

[0070] The GUI control unit 545a presents the data and information generated by the display control unit 545 to the user in a visually and easily operable format. For example, when generating AI-generated content, the GUI control unit 545a controls the user interface for inputting prompts for generation and uploading images and videos. Furthermore, the GUI control unit 545a also plays a role in calling appropriate application functions in response to user operations.

[0071] The location information acquisition unit 544 is a module that has the function of acquiring the current location information using the built-in GPS sensor of the smartphone 5 and communication information with wireless base stations. Specifically, the location information acquisition unit 544 acquires the location information using the Global Positioning System (GPS) using artificial satellites, base station positioning by triangulation based on radio wave strength and base station information from base stations, and Wi-Fi positioning using a database that combines Wi-Fi SSID (Service Set ID) and radio wave conditions with longitude and latitude.

[0072] Furthermore, in this embodiment, the location information acquisition unit 544 also includes a time information acquisition unit 544a. The location information and time information acquired by this location information acquisition unit 544 are linked as metadata, which is related data, when AI-generated content data is generated, and are stored as information about when and where the voice data was acquired. This function makes it possible to add temporal and geographical context to the voice data, enriching the story and background behind the voice data.

[0073] (3) Blockchain Platform 4 As described above, this embodiment provides a blockchain platform 4 that uses blockchain technology to guarantee the authenticity of AI-generated content. Specifically, as shown in Figure 6, the blockchain platform 4 is an information processing terminal that performs information recording (registration) and inquiry processing on the blockchain (distributed ledger system), which is a data storage means, and is composed of a server device and a database device. In this embodiment, the blockchain platform 4 includes a communication interface 43, an authentication unit 42, a tracking execution unit 44, an NFT transaction history database 41a, a key information database 41b, and an account database 41c.

[0074] The communication interface 43 is a module that transmits and receives data with other communication devices via the communication network 2, and in this embodiment, it is connected to each AI-generated content management server 3. The authentication unit 42 is a computer or software with that function that verifies the legitimacy of the accessor, and performs authentication processing based on the user ID that identifies each user. In this embodiment, the system obtains the user's unique public address, public key, user ID, password, etc. from the accessor's smartphone 5 via the communication network 2, and verifies whether the accessor has the right to access the information and whether the accessor is the person in question by comparing it with the key information database 41b.

[0075] The tracking execution unit 44 is a module that performs processing to realize traceability of AI-generated content using blockchain. To guarantee the authenticity of AI-generated content, it performs processing to query, manage, and update data on the blockchain, and realizes tracking and verification functions for AI-generated content. In this embodiment, the tracking execution unit 44 comprises a query processing execution unit 44a, a public address management unit 44b, an authenticity information verification unit 44c, and a data update unit 44d.

[0076] The inquiry processing execution unit 44a is a module that searches and queries tracking data recorded on the blockchain. For example, it refers to the traceability data of generated AI content, obtains information on specific content or transactions, and provides data necessary for authenticity verification and tracking.

[0077] The public address management unit 44b is a module that manages public addresses used on the blockchain, and performs processes to identify related users and entities by linking and matching public addresses.

[0078] The authenticity information verification unit 44c is a module that verifies whether AI-generated content is authentic, checking whether the content has been tampered with and whether it matches the data at the time of generation. Specifically, it compares the traceability data included in the audio watermark of the AI-generated content with the traceability data recorded on the blockchain, and verifies the consistency of the content and whether it has been tampered with based on the degree of agreement.

[0079] The data update unit 44d is a module that adds or modifies data on the blockchain. For example, when new AI-generated content is created, it records the transaction on the blockchain and updates existing data with the latest information, ensuring that the content generation history and authenticity information are always up-to-date.

[0080] Furthermore, the tracking execution unit 44 has a function to cooperate with the guarantee system cooperation unit 36c on the AI-generated content management server 3. This guarantee system cooperation function handles the management of traceability data related to AI-generated content, as well as security management, transaction records, service history storage, and other rights management processes, when requested by devices of other service organizations, such as the guarantee system cooperation unit 36c of the AI-generated content management server 3. Traceability processing is carried out through communication with these devices.

[0081] The public address management unit 44b is a module that manages users' public accounts (public addresses). It has a private key issuance function that issues public addresses generated from public keys in public-key cryptography to identify specific users, and private keys that, when paired with the public key, can identify the public key and are used for digital signatures in the management of NFT transactions of AI-generated content via public addresses. These issued public addresses and related key information are stored in the key information database 41b. When a user conducts an NFT transaction related to AI-generated content, this public address becomes the key that proves the legitimacy of the transaction. The public address functions as a user identifier on the blockchain and is used to uniquely identify the NFTs and transaction history owned by that user.

[0082] The data update unit 44d is a module that updates information related to AI-generated content, such as traceability data. It updates the blockchain when a user wants to change information about AI-generated content or when new AI-generated content is acquired. The data update unit adds the new information to the blockchain and performs procedures to maintain consistency with the previous information. For example, the data update unit 44d can acquire information related to AI-generated content via an NFT, add a public address related to the new NFT owner, and change the owner of the private key certified by the distributed ledger, thereby transferring ownership of the AI-generated content.

[0083] (How to manage AI-generated content) The AI-generated content management system 1 described above can be operated to implement the AI-generated content management method of the present invention. Figure 8 is a sequence diagram showing the operation of the AI-generated content management system. Note that the processing procedure described below is merely an example, and each process may be modified as much as possible. Furthermore, depending on the embodiment, steps in the processing procedure described below can be omitted, replaced, and added as appropriate.

[0084] First, user U1, who wishes to use the AI-generated content generation service, accesses the AI-generated content management server 3 from their smartphone 5 and initiates the authentication process (S201). Specifically, user U1 operates the AI-generated content management application on their smartphone 5 and attempts to access the AI-generated content management server 3. If the connection with the AI-generated content management server 3 is successful, the user is prompted to enter their user ID and password as login information. Once the user accurately enters this information, the server compares it with the authentication information stored in the user database 35a. If the authentication information matches, the user's authentication is successful and the user is identified (S101).

[0085] On the user terminal side, if the authentication process is successful, the AI ​​generation operation unit 546 can perform content generation operations on the generation system AI 6. Here, first, traceability data is input for this AI generation (S202), and traceability data is generated. In detail, on the smartphone 5 side, if the authentication process is successful, the AI ​​generation operation unit 546 inputs items as shown in Figure 10 to generate traceability data. Here, when user U1 performs an operation such as inputting text on the application, a process to request registration of traceability data is executed in cooperation with the traceability management unit 36 ​​on the AI ​​generation content management server 3 (S203), and the input information is saved (registered) in the traceability database 35b (S102).

[0086] Next, by performing an operation for AI content generation, a request (generation request) is sent from the user terminal to the AI ​​content generation management server 3 (S204). Upon receiving this request (S103), the AI ​​content generation is executed via the generation processing execution unit 36a (S104). This request includes necessary generation parameters (types such as text, images, and audio, desired style and format), such as prompts. The generation processing execution unit 36a converts the request received from the user into a form suitable for processing by the AI ​​model, and also performs pre-processing such as formatting the input data, noise reduction, and extraction of necessary information.

[0087] After this preprocessing, content is generated by the trained AI model. At this time, an audio watermark is inserted by the authentication data insertion processing unit 36d (S105). In this embodiment, at the same time that the AI-generated content is generated in the generation processing execution unit 36a, identification information unique to the AI-generated content is inserted as an audio watermark into the acoustic signal contained in the generated AI-generated content. This unique identification information is a unique recognition ID unique to the AI-generated content and the audio watermark, calculated by the generation processing execution unit 36a. The audio watermark is inserted into the audio data by encrypting this unique recognition ID using changes in the frequency, waveform, and amplitude of an inaudible sound. In this embodiment, identification information unique to the AI-generated content is inserted as an audio watermark, but it is also possible to use watermarks or icons in the form of images or videos.

[0088] Next, the system connects to the blockchain platform 4 and registers the original data of the generated AI-generated content as an NFT (S106), while simultaneously registering the traceability data (tracking data) to the blockchain (S107). In conjunction with this, the traceability data D1 is also registered in the traceability database 35b of the AI-generated content management server 3. When registering the traceability data D1 to this traceability database 35b,

[0089] Here, we will describe in detail the mechanism of the distributed ledger system in the blockchain platform 4 (guarantee system) mentioned above. In this embodiment, the blockchain platform 4 provides a blockchain interface service that connects to the blockchain. This service includes multiple nodes that store at least some or all of the data (traceability data and AI-generated content) generated by the AI-generated content management server 3. These nodes aggregate the stored data at predetermined timings, block it, and use the blocks to form a blockchain. The blockchain is then shared among the multiple nodes and stored as a distributed ledger.

[0090] Specifically, as shown in Figure 9, the AI-generated content management system 1 according to this embodiment issues a key pair of a public key PKa and a private key SKa based on public-key cryptography through the blockchain platform 4 when registering and retrieving traceability data related to AI-generated content, issuing, transferring, and canceling NFTs, etc., and generates a public address PAa from the public key PKa corresponding to the generated AI-generated content. This public address PAa is used as an address indicating the creator (distributor) of the AI-generated content, while the private key SKa is used for the digital signature when registering the AI-generated content to the public address PAa.

[0091] In this embodiment, the management of AI-generated content and traceability data is performed among numerous nodes on a P2P (Peer-to-Peer) network 90, and the tracking information is broadcast and shared to each node 90a to 90f within the P2P network 90. ​​As a result, a transaction history database (so-called blockchain) using a distributed ledger system is formed on the P2P network 90, and information regarding the generation and tracking of various tokens and AI-generated content, as well as transaction history related to NFTs, is stored.

[0092] In this embodiment, transactions using this distributed ledger system are processed, approved, and managed through the AI-generated content management server 3, including the registration of AI-generated content and its traceability data, the issuance of NFTs, and the rewriting of their owners. The registration and rewriting of this data involves generating a unique public address PAa for each AI-generated content, and performing registration at the time of content generation and inquiry at the time of content viewing through the public address PAa.

[0093] Then, each user (generator and viewer) uses the AI-generated content management system 1 to distribute the generated AI-generated content, and ensures the traceability of the content by allowing viewers to query tracking data via the public address PAa.

[0094] Here, we will describe in detail the process for querying tracking data. Figure 9 shows the processing procedure for registering AI-generated content and querying its traceability in the system according to this embodiment. In this embodiment, the account processing related to the registration of AI-generated content and querying its traceability utilizes the mechanism of the distributed ledger system according to this embodiment.

[0095] In this embodiment, traceability data D1 of the generated AI content is registered to the public address PAa through the blockchain platform 4, and the NFT of the AI-generated content data D2 is also registered to this public address PAa. A pair of private keys SKa corresponding to this address is formed for this public address PAa. Specifically, a private key SKa associated with a unique public address is generated using public-key cryptography. Next, a public key PKa is generated from the private key SKa based on an electronic signature algorithm such as Elliptic Curve Digital Signature Algorithm (ESDSA). The generated public key PKa and private key SKa form a key pair in public-key cryptography, and due to the nature of this public-key cryptography, it is possible to generate a public key PKa from a private key SKa, but it is impossible to generate a private key SKa from a public key PKa from a public key PKa due to computational complexity.

[0096] Next, the traceability data D1 and AI-generated content data D2 are registered with the node, linked to this public address PAa. As a result, the traceability data D1 and AI-generated content data D2 are linked to the public address PAa and made public. This traceability data D1 and AI-generated content data D2 can be freely viewed by anyone who has obtained the public key PKa related to the public address PAa. As a result, tracking data such as the creator and creation date of the AI-generated content can always be checked, and anyone can verify actions such as tampering with the voice watermark.

[0097] According to the blockchain mechanism, the tracking history of AI-generated content data, including its creation and transfer, is defined as a chain of digital signatures. When AI-generated content is transferred to another owner, the hash value of the previous transaction and the hash value of the next owner's public key are digitally signed with the owner's private key and added to the block history. This makes it possible to determine whether the transfer from the previous owner to the new owner has been tampered with.

[0098] In this way, traceability data D1 and AI-generated content data D2 are registered using the blockchain mechanism. Then, as a generation completion process (S108), post-processing is performed to check whether the generated AI content is appropriate and maintains the specified quality, and to make adjustments such as filtering and formatting as necessary. After that, the generated content is obtained by the user through processes such as downloading the completed content to the smartphone 5 or uploading it to the user's unique storage area (S205). At this time, the processing status of the request (success, failure, etc.) is also provided by the HTTP response status code.

[0099] Next, when the AI-generated content data is acquired on the smartphone 5 through user operation, completion processing is performed, such as distributing the AI-generated content to the user U1's personal My Page or other video distribution services provided by the AI-generated content management server 3 (S206).

[0100] (Effects / Actions) As described above, according to this embodiment, a unique audio watermark is attached to the AI-generated content, and the authentication information (content ID) that identifies the audio watermark is associated with traceability data D1, which includes information about the content's creation (who created it, when, and for what purpose, etc.), and stored on the blockchain. This system makes it possible to verify the origin and creation process of AI-generated content afterward, thereby ensuring the authenticity of AI-generated content and the transparency of its origin. As a result, it is expected that fraudulent content such as deepfakes will be eliminated from the market, and highly reliable content generation technologies will become widespread.

[0101] In particular, in this embodiment, a unique watermark (audio fingerprint) is inserted simultaneously with content generation, making it possible to more reliably track the origin and history of the generated audio. This allows for accurate identification of the source and clarification of responsibility even if the generated audio is misused by a third party.

[0102] Furthermore, according to this embodiment, by querying the traceability data D1 linked to the audio watermark, which is simultaneously added when content is generated, the entire history of the generated content can be tracked. Companies and users utilizing AI generation technology can verify how the generated audio was used and in which applications or solutions it was used, thereby proving the authenticity of the audio. In addition, by providing generation and querying as a set, the reliability of the audio generation technology is dramatically improved.

[0103] This embodiment makes it possible to establish traceability in AI-generated content, reduce the risk of fake voices circulating in the market, and provide an environment in which companies and individuals can safely utilize AI voice generation technology. This increases the reliability of the voice generation market, allowing large corporations and organizations to use AI-generated content with confidence.

[0104] It should be noted that the present invention is not limited to the embodiments described above, and the components can be modified and implemented in practice without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriately combining the multiple components disclosed in the above embodiments. For example, some components may be deleted from all the components shown in the embodiments. [Explanation of symbols]

[0105] C1…Video content D1…Traceability data (tracking data) D2... AI-generated content data PAa... Public Address PKa...Public Key SKa...Private Key U1…User (Generator) U2…User (viewer) 1…AI-generated content management system 2…Communication Network 3…AI-generated content management server 4… Blockchain platform 5(5a,5b)...Smartphone 6…Generation AI 22… Wireless base stations 31…Communication Interface 32…Location information management department 33…Authentication Department 34…Content Distribution Department 35a...User Database 35b...Traceability Database 35c... AI-generated content database 36…Traceability Management Department 36a...Generation processing execution unit 36b...Tracking Information Inquiry Department 36c...Warranty System Integration Department 36d... Authentication data insertion processing unit 41a...NFT Transaction History Database 41b…Key Information Database 41c…Account Database 42…Authentication Department 43…Communication Interface 44... Tracking Execution Unit 44a... Inquiry Processing Execution Unit 44b…Public Address Management Department 44c…Authenticity Information Verification Department 44d...Data update section 51…Communication Interface 52…Input Interface 52a...Touch panel 53…Output Interface 53a...Display section 54…Application execution unit 55...Memory 90…P2P network 90a~90f…Nodes 540…Content Viewing Department 541...Data extraction unit 542…Authenticity Inquiry Department 543...Synchronization Processing Unit 544...Location information acquisition unit 544a...Time information acquisition unit 545...Display Control Unit 545a...GUI Control Unit 546…AI generation operation section

Claims

1. A system for managing content generated based on information learned by artificial intelligence as AI-generated content, When generating the AI-generated content, an audio watermark insertion unit inserts unique identification information for the AI-generated content as an audio watermark into the audio signal included in the AI-generated content. A tracking data acquisition unit acquires tracking data including information related to the generation of the AI-generated content, A data storage means that stores at least a portion of the authentication information that identifies the aforementioned voice watermark in association with the aforementioned tracking data, When playing the AI-generated content, the inquiry unit extracts the audio watermark inserted within the AI-generated content being played, and queries the tracking data from the data storage means based on the extracted audio watermark. An AI-generated content management system characterized by having the following features.

2. The AI-generated content management system according to claim 1, characterized in that the tracking data acquisition unit acquires identifiers and variables defined for each item as tracking data related to the generation of the AI-generated content.

3. The aforementioned audio watermark is inserted as an inaudible sound within the acoustic signal included in the AI-generated content. The inquiry unit obtains the identification information by extracting the inaudible sound. The AI-generated content management system according to feature 1.

4. The data storage means comprises a plurality of nodes that encrypt and store at least a portion of the tracking data, The node aggregates at least a portion of the tracking data at a predetermined time and blocks it, concatenates this block with an existing block to form a blockchain, and shares the blockchain among multiple nodes to store it as a distributed ledger. The AI-generated content management system according to feature 1.

5. A method for managing content generated based on information learned by artificial intelligence as AI-generated content, When generating the AI-generated content, the audio watermark insertion unit inserts an identification information unique to the AI-generated content as an audio watermark into the acoustic signal contained in the AI-generated content. A data storage step in which tracking data including information relating to the generation of the AI-generated content is acquired, and at least a portion of the authentication information identifying the voice watermark is stored in association with the tracking data in the data storage means, When playing the AI-generated content, the inquiry unit extracts the audio watermark inserted within the AI-generated content being played, and queries the tracking data from the data storage means based on the extracted audio watermark in an inquiry step. A method for managing AI-generated content, characterized by including the following.

6. The AI-generated content management method according to claim 5, characterized in that, in the data storage step, identifiers and variables defined for each item related to the generation of the AI-generated content are acquired as the tracking data.

7. The aforementioned audio watermark is inserted as an inaudible sound within the acoustic signal included in the AI-generated content. In the query step, the identification information is obtained by extracting the inaudible sound. The AI-generated content management method according to feature 5.

8. The data storage means comprises a plurality of nodes that encrypt and store at least a portion of the tracking data, The node aggregates at least a portion of the tracking data at a predetermined time and blocks it, concatenates this block with an existing block to form a blockchain, and shares the blockchain among multiple nodes to store it as a distributed ledger. The AI-generated content management method according to feature 5.

9. A program for managing content generated based on information learned by artificial intelligence as AI-generated content, and which uses a computer, When generating the AI-generated content, an audio watermark insertion unit inserts unique identification information for the AI-generated content as an audio watermark into the audio signal included in the AI-generated content. A tracking data acquisition unit acquires tracking data including information related to the generation of the AI-generated content, A data storage means that stores at least a portion of the authentication information that identifies the aforementioned voice watermark in association with the aforementioned tracking data, When playing the AI-generated content, the inquiry unit extracts the audio watermark inserted within the AI-generated content being played and queries the tracking data from the data storage means based on the extracted audio watermark. An AI-generated content management system characterized by functioning as such.

10. The AI-generated content management program according to claim 9, characterized in that the tracking data acquisition unit acquires identifiers and variables defined for each item as tracking data related to the generation of the AI-generated content.

11. The aforementioned audio watermark is inserted as an inaudible sound within the acoustic signal included in the AI-generated content. The inquiry unit obtains the identification information by extracting the inaudible sound. The AI-generated content management program according to feature 9.

12. The data storage means comprises a plurality of nodes that encrypt and store at least a portion of the tracking data, The node aggregates at least a portion of the tracking data at a predetermined time and blocks it, concatenates this block with an existing block to form a blockchain, and shares the blockchain among multiple nodes to store it as a distributed ledger. The AI-generated content management program according to feature 9.

Images