Learning data right management system

Abstract

The present invention addresses the problem of providing a rights management system for generating learning data for machine learning models such as AI. [Solution] A right management system according to an embodiment assigns a learning data ID to learning data of a generative AI model for generating new content or new data, and accepts registration of right information of the learning data to which the learning data ID has been assigned. Registration of a generative AI model generated by using learning data permitted to be used is accepted from a model developer, and model registration information associated with a learning data ID of the used learning data is generated. A usage permission application for the registered generative AI model is accepted from the operator, and a usage fee corresponding to the right information of the learning data associated with the generative AI model permitted to be used is calculated for the operator who obtains the operating income by the content or data generated by the registered generative AI model. The use fees deposited from the operator are aggregated, and allocated amounts are calculated for each of the learning data providers.

Description

Technical Field

[0001] Embodiments of the present invention relate to rights management technology for learning data used in machine learning models such as generative AI models. Background Technology

[0002] The generation of machine learning models, such as AI models, and the use of machine learning models in business operations are booming. In particular, in recent years, technologies that use artificial intelligence, known as generative AI, to generate new content and new data have attracted attention.

[0003] Generative AI can learn patterns and generate new content and data based on learning data. This generative AI is a machine learning model built using machine learning methods such as deep learning.

[0004] Existing technical documents: Non-patent literature: Non-Patent Document 1: Publisher: National Diet Library, Bureau of Research and Legislative Examination; Author: Takashi Amemiya, Economic and Industrial Division; Title: "NFT Trends and Issues - Contains Holda - Protection and Consumer Protection at its Center"; Publication Date: April 4, 2023 (No. 1232); URL: https: / / dl.ndl.go.jp / view / prepareDownload?itemId=info:ndljp / pid / 12767935 Summary of the Invention

[0005] The problem that the invention aims to solve The aim is to provide a rights management system for learning data.

[0006] Methods for solving problems The rights management system implemented in this embodiment manages the rights of learning data used to generate generative AI models containing new content or data. The rights management system comprises: a learning data management department, which assigns learning data IDs to learning data providers, accepts registration of rights information for learning data with assigned learning data IDs, and accepts licensing applications from model developers for the learning data; a model management department, which accepts registration of generative AI models generated using licensed learning data, generates model registration information associated with the learning data IDs of the used learning data, and accepts licensing applications from operators for the registered generative AI models; and a usage fee management department, which calculates usage fees corresponding to the rights information of the licensed generative AI models for operators who generate revenue from content or data generated by the generative AI models, summarizes usage fees received from operators, and calculates and distributes amounts separately according to the learning data providers. Attached Figure Description

[0007] Figure 1 This is a schematic diagram illustrating the rights management system of the first embodiment.

[0008] Figure 2 This is a diagram showing the system structure and functional blocks of the rights management system of the first embodiment.

[0009] Figure 3 This is a diagram illustrating the processing flow centered on the learning data provider in the rights management system of the first embodiment.

[0010] Figure 4 This is a diagram illustrating the model developer-centric processing flow in the rights management system of the first embodiment.

[0011] Figure 5 This is a diagram illustrating the operator-centric processing flow in the rights management system of the first embodiment.

[0012] Figure 6 This is a diagram illustrating the allocation process and allocation mechanism to learning data providers in the rights management system of the first embodiment. Detailed Implementation

[0013] The embodiments will now be described with reference to the accompanying drawings.

[0014] Machine learning models are generated by processing learning data. Machine learning models can be broadly classified into two types.

[0015] The first type of machine learning model is an AI model that organizes or classifies information for prediction, inspection, etc. Examples include image recognition models and sound recognition models. The characteristic of this first type of machine learning model is that it automatically performs the determined actions and outputs appropriate responses, such as predictions and inspection results, based on the learned data.

[0016] On the other hand, the second type of machine learning model is generative AI models that generate (create) content and data. These include various types such as text-generating AI models, sound-generating AI models, image-generating AI models, and motion-generating AI models. The characteristic of generative AI models is that they generate new content such as articles (text), sounds, music, images, and motion-generating images.

[0017] Furthermore, in recent years, the emergence of generative AI has led to a surge of opinions and discussions on how to handle learning data. For example, voice-generating AI, which learns from the voice data of voice actors (actors, narrators) appearing solely in audio and video works, can generate new audio content based on the voice actors' voice data. Regardless of whether the discussion revolves around human-created content or AI-generated content, the rights holders of the learning data and the benefits derived from using their own voice data to create voice-generated AI are undeniably connected.

[0018] Thus, the new content and data generated by generative AI depend on the learning data used to build the generative AI. Therefore, the disorderly use of learning data is not appropriate from the perspective of protecting the rights of those who own the learning data. Therefore, the urgent task is to create platforms that can protect the rights of those who own the learning data; in other words, platforms that can effectively connect stakeholders when using generative AI to create new content or businesses.

[0019] The rights management system of the implementation provides a rights management platform that uniformly manages learning data providers, model developers, and operators who use generative AI models, and realizes a mechanism for distributing benefits based on the learning data of generative AI models that contribute to business revenue.

[0020] Furthermore, from another perspective, businesses using generative AI models that utilize learning data of unknown origin may face risks. For example, licensing is sometimes required when copyrighted works use learning data. Therefore, clearly understanding the type of learning data used to develop the generative AI model can mitigate these risks and allow for the promotion of the quality of the learning data and the value of the generative AI model. Additionally, model developers can clearly communicate the source of the learning data to businesses, enabling them to promote the model to potential buyers. A rights management platform can facilitate the development of generative AI models and enhance the value of businesses using these models.

[0021] (First Implementation) Figures 1 to 6 This is a diagram used to illustrate the first embodiment.

[0022] Figure 1 This is a schematic diagram illustrating a rights management system for learning data of machine learning models that include generative AI models. The rights management system includes a management device 100. The management device 100 provides a rights management platform for use by learning data providers, model developers, and operators.

[0023] In the following description, a generative AI model will be used as an example of a machine learning model. The learning data includes text data, audio data, music data, image data, and motion picture data, etc., which are used for generative AI models. Regardless of the category or content of the learning data, the rights management platform of this embodiment performs rights management on the learning data used for learning processing (model generation processing) of the generative AI model.

[0024] Those who have rights to learning data are called learning data providers. For example, in the case of audio data, the learning data provider is the individual who provides the audio or the company to which that individual belongs. For example, an individual such as a voice actor or actor, or the company to which a voice actor or actor belongs, are equivalent to this.

[0025] Model developers are developers of machine learning models that use learning data provided by learning data providers to build and generate generative AI models. There are no particular restrictions on the types or uses of generative AI models registered on the rights management platform, as long as they use learning data registered on the platform. Furthermore, model developers can not only develop generative AI models but also develop computer systems incorporating such models and provide them to operators. Model developers can be individuals, companies, or other organizations.

[0026] An operator is someone who operates a business by receiving generative AI models from model developers. For example, an operator may purchase a generative AI model from a model developer, or purchase the right to use a computer system containing a generative AI model provided by the model developer, or provide content or data generated by the generative AI model, or provide services utilizing content or data generated by the generative AI model. Alternatively, an operator may enter into a transaction contract with a model developer to commission the development of a generative AI model. An operator is an individual or company that operates a business for profit by managing or operating generative AI models or computer systems containing generative AI models.

[0027] Data providers, model developers, and operators register users as a preparation for utilizing the rights management system (management device 100).

[0028] The learning data provider registers the learning data in the rights management system. The rights management system identifies the registered learning data. Specifically, it assigns a learning data ID to each learning data item and processes the registration of rights information for the learning data assigned the learning data ID. Through the identification of the learning data and the registration of rights information, the licensing of the learning data is completed.

[0029] Furthermore, learning data can be registered as multiple data groups, i.e., datasets. When a dataset consisting of multiple learning data groups is registered, learning data IDs are assigned to the dataset and the multiple learning data points contained within it in a recognizable manner. Alternatively, the registration of learning data can also be controlled to be a single learning data point, rather than a data group.

[0030] Model developers can utilize the rights management system to retrieve learning data registered therein, view accompanying rights information, and apply for usage licenses. The rights management system processes usage license applications from model developers. Model developers can then obtain the licensed learning data from the rights management system.

[0031] Model developers use the acquired learning data to generate generative AI models. These models are then registered with the rights management system. The rights management system processes registrations and generates model registration information that links the generated model to the learning data ID used in the learning data. In other words, it establishes a connection between the generated generative AI model and the learning data used, thus making the generative AI model identifiable. This completes the preparation for licensing the generative AI model.

[0032] Furthermore, model developers can obtain learning data in two ways: a first scenario, by downloading and obtaining learning data from the management device 100 through the rights management system; and a second scenario, by obtaining learning data without going through the rights management system. In the second scenario, for example, learning data can be obtained directly from a learning data provider.

[0033] In the first scenario, since the learning data is downloaded from the rights management system, the generative AI model is generated using learning data with a license. In the second scenario, since the learning data is not provided through the rights management system, sometimes learning data that is not licensed for use on the rights management system is used to generate the generative AI model. In this case, the model developer can later apply for a license to use the learning data used in generating the generative AI model. For example, it can be configured to retrieve the learning data used when registering the generative AI model and associate the learning data registered with the rights management system with the generative AI model. Furthermore, even in the second scenario, it is possible to access the rights management system and apply for a license to use the corresponding learning data before generating the generative AI model using the learning data. In this case, the licensed learning data is used to generate the generative AI model.

[0034] The transaction takes place separately between the operator and the model developer. The operator obtains the desired generative AI model from the model developer, or the right to use a computer system containing the generative AI model provided by the model developer. The operator applies for a license to use the generative AI model used in the business with the rights management system. The rights management system processes the license application from the operator for the registered generative AI model and establishes a link between the operator and the generative AI model. This establishes a link between the business and learning data using the generative AI model.

[0035] Here, we explain the various licenses for using learning data and generative AI models. The licensing of learning data is the process of establishing a connection between the learning data and the model developer. That is, the license records which learning data the model developer has used. The form of the license is arbitrary; for example, it can be a structure of a publicly known license agreement, or it can be a structure of learning data usage registered in a rights management system for model developer certification (license). Therefore, applying for a learning data license is the act of a model developer applying to use (using) the learning data registered in the rights management system. The process of establishing a connection between the model developer applying to use the learning data and the corresponding learning data (learning data provider) is the licensing process.

[0036] Similarly, licensing generative AI models is the process of establishing a connection between model developers and operators, and it's used to record the history of which generative AI models an operator uses in their business operations. The form of a generative AI model license is also arbitrary; for example, it can apply a structure based on a well-known licensing agreement. Therefore, a generative AI model license application refers to the act of an operator applying to use (or using) a generative AI model registered in the rights management system. The process of establishing a connection between the operator applying to use the generative AI model and the corresponding generative AI model is the licensing process.

[0037] Next, the rights management system calculates the usage fee payable by the operator based on the rights information of the learning data associated with the licensed generative AI model, and processes the operator's usage fee request. At this time, the operator registers (reports) business information (revenue-related information) containing business revenue related to the licensed generative AI model to the rights management system. The rights management system can calculate the usage fee requested from the operator based on the declared business revenue. The operator pays the requested usage fee. The rights management system aggregates the usage fees received from the operator and calculates the allocation amount separately for each learning data provider. The calculated allocation amount is deposited into the learning data provider's transfer destination.

[0038] In this way, the rights management system can provide a unified platform for managing the rights of learning data providers, model developers, and operators, and realize a mechanism for distributing the benefits of learning data from generative AI models that contribute to business revenue.

[0039] Figure 2 This diagram illustrates the system configuration and functional blocks of the rights management system according to this embodiment. The rights management system is configured to include a management device 100. The management device 100 includes a communication device 110, a control device 120, and a storage unit 130. Learning data providers, model developers, and operators connect to the management device 100 from various user terminals via networks such as IP networks and use the rights management platform.

[0040] The control device 120 is configured to include a user interface control unit 121, a user management unit 122, a learning data management unit 123, a model management unit 124, and a usage fee management unit 125.

[0041] The user interface control unit 121 is a processing unit that provides various screens for utilizing the rights management system to the user terminals of learning data providers, model developers, and operators, and performs screen control corresponding to the operation of the user terminals.

[0042] User Management Unit 122 provides user registration functionality via a user registration screen. User Management Unit 122 includes Registration Control Unit 1221. Registration Control Unit 1221 stores user information entered by learning data providers through the learning data provider registration screen in storage device 130, categorized by learning data provider. The user information of the learning data provider includes Learning Data Provider ID, login password, registrant information (name, address, contact destination, email address, etc.), organization information (organization name, address, contact destination, email address, etc.), registration category (indicating the category of the learning data provider), and transfer destination information (user fee transfer destination information, etc.).

[0043] Then, the registration control unit 1221 stores the user information entered by the model developer through the model developer registration screen into the storage device 130 according to the model developer. The user information of the model developer includes the model developer ID, login password, registrant information (name, address, contact destination, email address, etc.), organization information (organization name, address, contact destination, email address, etc.), registration category (indicating the category of the model developer), etc.

[0044] In addition, the registration control unit 1221 stores the user information entered by the operator through the operator registration screen in the storage device 130 according to the operator. The operator's user information includes operator ID, login password, registrant information (name, address, contact destination, email address, etc.), organization information (organization name, address, contact destination, email address, etc.), registration category (indicating the operator's category), and usage fee settlement information (credit card information, account transfer information, etc.).

[0045] Next, the learning data management department 123 includes the learning data registration and control department 1231, the learning data rights information management department 1232, the learning data identity management department 1233, and the learning data usage license management department 1234.

[0046] The learning data registration and control unit 1231 assigns learning data IDs to the learning data uploaded to the management device 100 through the learning data registration screen and stores them in the storage device 130.

[0047] The Learning Data Rights Information Management Department 1232 controls the input of rights information through the user terminal of the learning data provider, and stores the input rights information in the storage device 130 in association with the learning data ID. The rights information may include whether a license is permitted, the scope of the license, information on prohibited content, and fee settings. The scope of the license specifies the purpose and use of the generative AI model generated from the learning data. For example, it can be set that the use of the generative AI model is permitted for non-profit purposes but not for profit purposes, or that such use is permitted for internal company education. Information on prohibited content specifies the scope of provision, age restrictions, etc., of content, data, or services provided using the generative AI model. For example, it can be set that the application of generative AI models in content, data, or services provided to children under 12 years of age is prohibited.

[0048] The fee setting information specifies the usage fee for the learning data. The fee setting is arbitrary; for example, it can be set as a usage fee corresponding to the amount of learning data used in the learning process of the generative AI model, a usage fee corresponding to the sales price of the generative AI model, or a usage fee relative to the revenue obtained by the operator from the use of the generative AI model, etc.

[0049] The Learning Data Identity Management Unit 1233 generates and processes identity verification information for learning data, and performs identity verification. For example, it uses a hash function based on a prescribed algorithm to generate a hash value for the learning data. This hash value is associated with a learning data ID and managed to verify (prove) that the learning data is genuine and has not been tampered with or edited. For example, in the event of an identity verification request from a model developer, the Learning Data Identity Management Unit 1233 can generate a hash value for the learning data received from the model developer's user terminal, compare it with a registered hash value, and provide the identity verification result to the model developer.

[0050] Furthermore, the learning data identity management unit 1233 can provide the model developer's user terminal with an identity check information generation program containing a hash function based on a prescribed algorithm. The model developer executes the provided identity check information generation program on the learning data to generate a hash value. The model developer sends the hash value generated from the user terminal to the management device 100. The learning data identity management unit 1233 compares the received hash value with the registered hash value to perform identity check processing and provides the result to the user terminal.

[0051] The learning data identity check function determines whether the learning data is registered in the rights management system, and it has the following implications: Model developers can verify the authenticity of the learning data; and model developers can verify whether learning data obtained from sources other than the rights management system is registered in the rights management system. As preparation for generating generative AI models, model developers can confirm in advance whether registered learning data exists. If the learning data obtained from outside the rights management system is registered, the model developer can apply for a license to use the rights management system.

[0052] The Learning Data Usage License Management Department 1234 processes usage licenses for learning data registered in the rights management system. The department accepts usage license applications for registered learning data from user terminals of model developers and processes them according to the learning data ID.

[0053] The Learning Data Management Department 123 manages the learning data attribute information uniformly according to the learning data, including the learning data ID, identity check information, learning data provider ID, rights information ID, and dataset description. Furthermore, the Learning Data Management Department 123 can control the licensed learning data to ensure it can be downloaded to the user terminals of model developers.

[0054] Furthermore, as mentioned above, there are cases where the learning data includes datasets consisting of multiple data points, but the datasets are large in size, making it sometimes difficult to upload them to the management device 100. Therefore, it is also possible to configure the system so that the learning data itself is not uploaded to the management device 100, but only the registration information of the learning data is input into the management device 100.

[0055] Specifically, the learning data management unit 123 controls the provision of the identity check information generation program managed by the learning data identity management unit 1233 to the user terminal of the learning data provider. Furthermore, the learning data provider executes the provided identity check information generation program on the learning data to generate a hash value. The learning data provider sends the generated hash value to the management device 100. The learning data registration control unit 1231 can assign a learning data ID to the received hash value and register it, generating learning data attribute information.

[0056] In this case, the model developer can obtain learning data through means other than the management device 100. For example, the model developer can obtain the learning data directly from the learning data provider or through a download site used by the learning data provider. If the model developer obtains the learning data through a path other than the management device 100, the model developer can access the management device 100, use the identity check function provided by the learning data identity management unit 1233 to perform an identity check on the obtained learning data, and apply for a license to use the learning data obtained through the license function provided by the learning data license management unit 1234.

[0057] The model management department 124 is composed of a model registration control department 1241, a model identity management department 1242, and a model usage license management department 1243.

[0058] The model registration control unit 1241 assigns model IDs to generative AI models generated using learning data registered in the management device 100, and generates model registration information that associates the learning data IDs of the learning data used with the model IDs. Specifically, model developers register generative AI models generated using learning data registered in the rights management system with the rights management system. The model registration control unit 1241 accepts model characteristic information and assigns model IDs through the generative AI model registration screen. Furthermore, along with the registration of the generative AI model, an identity check information generation program, including a hash function for generating identity check information, is provided to the user terminal. The model developer uses the source data of the generative AI model containing model characteristic information as the object and executes the identity check information generation program to generate a hash value. The generated hash value, along with the model characteristic information, is input through the generative AI model registration screen.

[0059] The model registration control unit 1241 generates model registration information containing model ID, model developer ID, model characteristic information and identity check information (hash value), and stores it in the storage device 130.

[0060] Here, the model characteristic information is explained. The model characteristic information includes the learning data ID and the quantity of data used in the learning process of the generative AI model. It can be configured to include any characteristic information of the generative AI model in addition to the learning data ID and the quantity of data. The model registration control unit 1241 can be configured to generate model registration information, which includes identity check information generated based on the source data of the generative AI model containing the model characteristic information. Thus, by including the learning data ID and the quantity of data used in the hash value used to check the identity of the generative AI model, even generative AI models using the same learning data ID will generate different hash values ​​(separate model registration information) if the quantity of data is different. The rights management system registers and manages the generative AI model while maintaining the inseparability of the learning data ID and its quantity.

[0061] Model Identity Management Unit 1242 provides the same functions as Learning Data Identity Management Unit 1233 of Learning Data Management Unit 123. That is, it provides the function to determine whether a generative AI model is a registered generative AI model in the rights management system, which is used by operators of generative AI models to confirm whether they are registered in the rights management system. As preparation for operating a business using a generative AI model, operators can confirm in advance whether the generative AI model is registered. If the generative AI model used by the operator is a registered generative AI model, the operator applies for a usage license from the rights management system.

[0062] The Model Usage License Management Department 1243 processes the licensing of generative AI models for operators. The Model Usage License Management Department 1243 receives license applications for registered generative AI models from the operator's user terminal and processes the licensing to associate the model ID with the operator.

[0063] The usage fee management department 125 is composed of an operation information receiving department 1251, a usage fee calculation department 1252, a usage fee payment management department 1253, and a usage fee allocation management department 1254.

[0064] The Business Information Receiving Department 1251 provides the function of receiving revenue-related information from the business operator's user terminal regarding the use of the licensed generative AI model. The Usage Fee Calculation Department 1252 calculates the usage fee corresponding to the rights information associated with the learning data based on the revenue-related information. Specifically, it calculates the usage fee paid by the business operator based on the rate setting information included in the rights information and the business revenue from using the generative AI model included in the revenue-related information. The Usage Fee Payment Management Department 1253 processes requests for usage fee payments from the business operator. Request processing involves requesting payment of usage fees from the business operator. It may also include settlement processing such as credit settlement or account transfer accompanying the request. The Usage Fee Allocation Management Department 1254 summarizes the usage fees received from the business operator and calculates the allocation amount separately for each learning data provider. The Usage Fee Management Department 125 processes the transfer of the calculated allocation amount to the learning data provider's designated transfer destination.

[0065] Storage device 130 stores user information 1301, learning data 1302, rights information 1303, learning data attribute information 1304, learning data usage license information 1305, model registration information 1306, model usage license information 1307, business information 1308, usage fee collection information 1309, and usage fee allocation information 1310.

[0066] Figure 3 This is a diagram illustrating the processing flow centered on the learning data provider within the rights management system.

[0067] like Figure 3 As shown, the learning data provider accesses the management device 100 to register as a user (S301). The management device 100 assigns a learning data provider ID and processes the registration of the entered user information (S101). The learning data provider logs in through the user terminal based on the learning data provider ID (S302), and the management device 100 performs login authentication processing (S102).

[0068] After login authentication, the learning data provider registers the learning data (S303). The learning data provider uploads the learning data to the management device 100 via the user terminal. The management device 100 generates identity check information for the uploaded learning data (S103). The management device 100 assigns a learning data ID to the uploaded learning data and associates the learning data ID with the identity check information to save the learning data (S104). Along with the upload of the learning data, the learning data provider inputs the rights information of the registered learning data (S304). The management device 100 registers the input rights information (S105), assigns a rights information ID to the input rights information, and associates the rights information ID with the learning data ID. The rights information associated with the learning data ID is saved (S106). The management device 100 generates learning data attribute information through the learning data registration process and the rights information registration process, and stores it in the storage device 130.

[0069] Figure 4 This is a diagram illustrating the developer-centric processing flow within the rights management system.

[0070] like Figure 4 As shown, the model developer accesses the management device 100 to register as a user (S401). The management device 100 assigns a model developer ID and processes the registration of the entered user information (S111). The model developer logs in through the user terminal based on the model developer ID (S402), and the management device 100 performs login authentication processing (S112).

[0071] After logging in and authenticating, the model developer is able to retrieve learning data (S403). The management device 100 performs retrieval processing based on the retrieval conditions specified by the model developer and provides the retrieval results to the model developer's user terminal (S113).

[0072] The management device 100 controls the accessibility of the retrieved learning data rights information (S114), and the model developer confirms the rights information through the user terminal (S404).

[0073] The model developer applies for a license to use the desired learning data (S405). The management device 100 accepts the license application and processes it, associating the requested learning data with the model developer (S115). The management device 100 controls the licensed learning data to be downloadable, and based on the learning data acquisition request from the model developer (S406), provides the learning data to the model developer's user terminal (S116).

[0074] The model developer develops and generates a generative AI model using the learning data registered in the management device 100 (S407). The model developer registers the generated generative AI model in the management device 100. At this time, the model developer records the learning data ID and the number of data points of the learning data used in the model characteristic information of the generated generative AI model.

[0075] The model developer obtains the identity check information generation program from the management device 100 (S409, S117), uses the source data, which includes model characteristic information that records the learning data ID and the number of data, as the object, and executes the identity check information generation program on the user terminal to generate identity check information (S410).

[0076] The model developer inputs identity check information and model characteristic information through the model registration screen (S411). The management device 100 assigns a model ID, generates model registration information containing identity check information and model characteristic information (S118), and saves it (S119).

[0077] Figure 5 This is a diagram illustrating the operator-centric processing flow within the rights management system.

[0078] like Figure 5 As shown, the operator accesses the management device 100 to register as a user (S501). The management device 100 assigns an operator ID and processes the registration of the entered user information (S131). The operator logs in through a user terminal based on the operator ID (S502), and the management device 100 performs login authentication processing (S132).

[0079] After login authentication, the operator can retrieve generative AI models (S503). The management device 100 performs retrieval processing based on the retrieval conditions specified by the operator and provides the retrieval results to the operator's user terminal (S133).

[0080] The operator applies for a license to use the desired generative AI model (S404). The management device 100 accepts the license application and generates usage fee information based on the rights information of the learning data associated with the generative AI model (S134). The management device 100 provides the generated usage fee information to the operator's user terminal, and the operator expresses their consent to the license through the user terminal (S505). Based on the operator's consent to the license, the management device 100 processes the license application and associates the applied generative AI model with the operator (S135).

[0081] User fee information, for example, is generated based on rights information of learning data associated with generative AI models, and is distinguished by operator ID and model ID. User fee information can consist of revenue-related information registered by the operator, user fee request history, and user fee payment history (deposit history).

[0082] The operator uses the registered machine learning model in its operations and generates revenue-related information concerning the usage fee information (S506). The operator registers (reports) the generated revenue-related information to the management device 100 (S507).

[0083] The management device 100 calculates the usage fee corresponding to the revenue-related information based on the usage fee charging information (S136). The management device 100 processes the request for the calculated usage fee (S137), and the operator confirms the request content (S508).

[0084] The operator uses the operator of the operation management device 100 (rights management platform) as the transfer destination to process the receipt of the requested usage fee (S509). Alternatively, as mentioned above, known methods such as credit card settlement or account transfer can also be appropriately applied to process the operator's usage fee receipt. The management device 100 confirms the receipt of the usage fee from the operator (S138), generates a receipt history (payment history), and updates the usage fee charging information (S139).

[0085] Figure 6 This diagram illustrates the allocation process and mechanism for learning data providers within the rights management system.

[0086] like Figure 6 As shown, the management device 100 performs usage fee allocation processing at certain intervals (S151). Specifically, the management device 100 summarizes usage fee collection information (usage fee collection log) (S152) and allocates the summarized usage fees according to the learning data IDs used in the generative AI model (S153). That is, the allocation amount is calculated separately according to the learning data ID. The management device 100 processes the payment of the calculated allocation amount to the corresponding learning data provider's transfer destination (S154). After the payment processing is completed, a notification of the deposit is sent to the user terminal of the learning data provider (S155).

[0087] The management device 100, acting as a rights management platform, identifies the learning data of learning data providers and establishes a connection between it and the generative AI model generated using the learning data. Furthermore, by linking operators who actually use the generative AI in their operations with the generative AI model, it is possible to link the operations conducted by operators using the generative AI model with the learning data provider. This allows for the creation of a path to distribute the operators' operating profits to the learning data provider.

[0088] In addition, such as Figure 6 As shown, sometimes the model developer and the operator belong to the same organization. In this case, the model developer and the operator register as users separately, and after obtaining the aforementioned learning data usage license and model usage license, the operator establishes a connection with the learning data provider.

[0089] The rights management system of this embodiment can provide a rights management platform that can effectively connect stakeholders when using generative AI to generate new content or services.

[0090] Through the rights management platform, it is possible to uniformly manage learning data providers, model developers, and operators who use generative AI models, and to distribute the benefits of learning data from generative AI models that contribute to business revenue to learning data providers.

[0091] The implementation method has been described above, but the rights management system is not limited to generative AI models; it can also be applied to the first machine learning model described above. That is to say, not only for the revenue generated from new content and data based on generative AI models, but also for the revenue generated from predictions and judgments made by previous machine learning models such as image recognition AI models and sound recognition AI models, the learning data can sometimes contribute to the machine learning model. Therefore, regardless of whether the machine learning model generates new content or new data, it can protect the rights holders of the learning data.

[0092] Furthermore, the various functions constituting the rights management system can be implemented by a single computer device or distributed across multiple computer devices. Additionally, the storage area for user terminals capable of storing and providing learning data to model developers can be separated from the rights management system. For example, the rights management system can also be constructed in conjunction with an external storage device that provides data services for data storage and download.

[0093] Alternatively, the learning data identity management unit 1233 and the model identity management unit 1242 can also be configured as an identity management unit, which provides functions for generating identity check information for both learning data and machine learning models containing generative AI models, providing identity check information generation procedures, and comparing with registered identity check information.

[0094] Furthermore, the various functions constituting the rights management system can be implemented through programs. The computer programs prepared in advance for implementing each function are stored in an auxiliary storage device. The control unit, such as the CPU, reads the program stored in the auxiliary storage device into the main storage device. The control unit executes the program read from the main storage device, thereby enabling the functions of each unit to operate.

[0095] Furthermore, the above-described program can also be provided to a computer in the form of a computer-readable recording medium. Examples of computer-readable recording media include CD-ROMs, Blu-ray Disc Rewritable, DVD-ROMs, MO (Magneto Optical) disks, floppy disks, hard disks, SD cards, and USB flash drives. Additionally, hardware devices such as integrated circuits (ROM, RAM, and other IC chips) specifically designed and configured for the purposes of this invention are also included as recording media.

[0096] Furthermore, including the aforementioned procedures, this invention is not limited to execution on the architecture of a von Neumann computer, but can also be executed on the architecture of a so-called non-von Neumann computer, such as a neural computer based on the structure of brain neural circuits or a quantum computer that applies quantum mechanics to information processing.

[0097] Furthermore, embodiments of the present invention have been described, but these embodiments are provided as examples and are not intended to limit the scope of the invention. This new embodiment can be implemented in various other ways, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments or variations thereof are included within the scope or spirit of the invention, and are included within the scope of the invention as described in the claims and its equivalents.

[0098] Explanation of reference numerals in the attached figures: 100 Management Device 110 Communication device 120 control device 121 User Interface Control Department 122 User Management Department 1221 Registration and Control Department 123 Learning Data Management Department 1231 Learning Data Registration and Control Department 1232 Learning Data Rights Information Management Department 1233 Learning Data Identity Management Department 1234 Learning Data Usage License Management Department 124 Model Management Department Model Registration Control Department 1241 Model Identity Management Department 1242 Model 1243 Licensing Management Department 125 Usage Fee Management Department 1251 Business Information Reception Department 1252 Usage Fee Calculation Department 1253 Fee Payment Management Department 1254 Usage Fee Allocation Management Department 130 storage devices 1301 User Information 1302 Learning Data 1303 Rights Information 1304 Learning Data Attribute Information 1305 Learning Data Usage License Information Model 1306 Registration Information Model 1307 License Information 1308 Business Information 1309 Usage Fee Information 1310 Usage Fee Allocation Information

Claims

1. A rights management system, which is a rights management system for learning data of a generative AI model, wherein the generative AI model generates new content or new data, characterized in that, The right management system has: a learning data management section that assigns a learning data ID to learning data of a learning data provider, accepts registration of right information of the learning data to which the learning data ID is assigned, and accepts, from a model developer, a use permission application for the learning data; a model management section that accepts registration of a generative AI model generated using the learning data for which use is permitted, generates model registration information in association with the learning data ID of the used learning data, and accepts, from a business operator, a use permission application for the registered generative AI model; and a usage fee management section that calculates, for a business operator who earns a business income through content or data generated by the generative AI model, a usage fee corresponding to the right information of the learning data in association with the generative AI model for which use is permitted, aggregates the usage fees from the business operators, and calculates an allocation amount for each learning data provider.

2. The right management system according to claim 1, wherein the learning data management section has: a learning data registration control section that assigns a learning data ID to learning data uploaded from a terminal of a learning data provider and registers the learning data ID to a prescribed storage area; a learning data right information management section that performs input control of right information through a terminal of the learning data provider and registers the input right information in association with the learning data ID to a prescribed storage area; and a learning data use permission management section that accepts, from a terminal of a model developer, a use permission application for the registered learning data and performs use permission processing for each learning data ID, the learning data management section performs control so that the learning data for which use is permitted can be downloaded to the terminal of the model developer.

3. The right management system according to claim 2, wherein the learning data management section has a learning data identity management section that, as an object of the uploaded learning data, executes an identity check information generation program based on a hash function constituted by a prescribed algorithm and generates identity check information, the learning data registration control section registers the learning data in association with the learning data ID and the identity check information.

4. The right management system according to claim 1, wherein the learning data management section has: a learning data identity management section that provides, to a terminal of a learning data provider, an identity check information generation program that is a program based on a hash function constituted by a prescribed algorithm; a learning data registration control section that accepts, from a terminal of a learning data provider, identity check information generated by executing the identity check information generation program as an object of the learning data, assigns a learning data ID, and registers the learning data ID to a prescribed storage area; a learning data right information management section that performs input control of right information through a terminal of the learning data provider and registers the input right information in association with the learning data ID to a prescribed storage area; and a learning data use permission management section that accepts, from a terminal of a model developer, a use permission application for the registered learning data and performs use permission processing for each learning data ID. ​ The learning data use license management section receives a use license application for the registered learning data from a terminal of the model developer, and performs use license processing for each learning data ID.

5. The right management system according to claim 3 or 4, wherein The learning data identity management section receives learning data or identity check information of the learning data from each terminal of the learning data provider, the model developer, and / or the operator, collates the identity check information generated by executing the identity check information generation program on the received learning data with the registered identity check information, and outputs an identity check result.

6. The right management system according to claim 1 or 2, wherein The model management section has: a model registration control section that assigns a model ID to each generative AI model generated using learning data, and generates model registration information including the learning data ID of the used learning data and the model ID; and a model use license management section that receives a use license application for the registered generative AI model from a terminal of the operator, and performs use license processing for each model ID.

7. The right management system according to claim 6, wherein The model management section further has a model identity management section that provides an identity check information generation program based on a hash function constituted by a prescribed algorithm to a terminal of the model developer, The model registration control section receives identity check information generated by executing the identity check information generation program on source data of the generative AI model from a terminal of the model developer, and generates model registration information including the learning data ID, the model ID, and the identity check information.

8. The right management system according to claim 7, wherein The generative AI model holds model characteristic information including the learning data ID of the used learning data and the number of data, The model registration control section receives identity check information generated on source data of the generative AI model including the model characteristic information from a terminal of the model developer, and generates model registration information.

9. The right management system according to claim 7, wherein The model identity management section receives identity check information generated on source data of the generative AI model from each terminal of the learning data provider, the model developer, and / or the operator, collates the received identity check information with the registered identity check information, and outputs an identity check result.

10. The right management system according to claim 1, wherein The usage fee management section has: an operation information reception section that receives, from a terminal of the operator, revenue association information associated with operation revenue obtained by content or data generated by the generative AI model permitted to be used; a usage fee calculation section that calculates, based on the revenue association information, a usage fee corresponding to right information of the learning data associated with the generative AI model permitted to be used; The usage fee payment management unit performs request processing of usage fees paid by the operator; and The usage fee distribution management unit aggregates usage fees paid in from the operator, calculates distribution amounts separately for the learning data providers, and pays the distribution amounts to the transfer destination of the learning data providers.

11. A program executed by a computer that performs right management of learning data for a generative AI model that generates new content or new data, the program implementing functions of: a first function that assigns a learning data ID to learning data of a learning data provider, accepts registration of right information of the learning data to which the learning data ID is assigned, and accepts a usage permission application for the learning data from a model developer; a second function that accepts registration of a generative AI model generated using the learning data for which usage is permitted, generates model registration information that associates with the learning data ID of the used learning data, and accepts a usage permission application for the registered generative AI model from an operator; and a third function that calculates, for an operator who earns an operating income from content or data generated by the generative AI model, a usage fee corresponding to the right information of the learning data associated with the generative AI model for which usage is permitted, aggregates usage fees paid in from the operator, and calculates distribution amounts separately for the learning data providers.

12. A right management system of learning data, characterized by comprising: having: a learning data management unit that assigns a learning data ID to learning data of a learning data provider, accepts registration of right information of the learning data to which the learning data ID is assigned, and accepts a usage permission application for the learning data from a model developer; a model management unit that accepts registration of a machine learning model generated using the learning data for which usage is permitted, generates model registration information that associates with the learning data ID of the used learning data, and accepts a usage permission application for the registered machine learning model from an operator; and a usage fee management unit that calculates, from the right information of the learning data associated with the machine learning model for which usage is permitted, a usage fee paid by the operator, aggregates usage fees paid in from the operator, and calculates distribution amounts separately for the learning data providers.

Images