A text generation type AI supervision method and system based on blockchain technology

Abstract

The application provides a text generation type AI supervision method based on a blockchain technology, and the method comprises the following steps: S110, constructing a supervision blockchain network; S120, deploying a supervision smart contract on the supervision blockchain network; S130, determining whether to authorize a user to use an AI model based on the supervision smart contract; S140, in response to the user being authorized to use the AI model, generating text information corresponding to the user; S150, storing the text information and user information into the supervision blockchain network; and S160, querying and analyzing the information stored on the supervision blockchain network. By constructing a blockchain network, designing and deploying a supervision special-purpose smart contract, storing user information, input data and AI output data onto the blockchain, and allowing the supervision department to query and trace, the problem that the prior art cannot efficiently supervise the content generated by the text generation type AI is solved, and the transparency and traceability of the generative AI are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a text generation type AI supervision method and system based on blockchain technology. BACKGROUND

[0002] With the wide application of text generation type AI such as ChatGPT, more and more text information on the Internet will be generated by artificial intelligence models rather than humans. However, although developers of artificial intelligence models have done a lot of work to avoid the generation of harmful and unhealthy information, users still try to input different prompts to guide the artificial intelligence model to generate harmful content. This large amount of harmful content generated by AI quickly poses new challenges to the supervision of the Internet.

[0003] Harmful text content generated by AI has the characteristics of rapidity, large quantity, and concealment compared to traditionally written text. Traditional Internet supervision methods are costly and inefficient, and cannot cope with this situation, and cannot trace the information publisher and the specific artificial intelligence model, which can easily cause social risks.

[0004] The above problems are urgent to be solved. SUMMARY

[0005] The purpose of the present application is to provide a text generation type AI supervision method and system based on blockchain technology to improve the existing technology and efficiently supervise and trace the content generated by text generation AI, thereby helping to strengthen the supervision of the Internet.

[0006] The technical scheme adopted by the present application to solve its technical problems is: on the one hand, the present application provides a text generation type AI supervision method based on blockchain technology, the method comprising: S110: constructing a supervision blockchain network; S120: deploying a supervision smart contract on the supervision blockchain network; S130: determining whether to authorize a user to use an AI model based on the supervision smart contract; S140: in response to the user being authorized to use the AI model, generating text information corresponding thereto; S150: storing the text information and user information into the supervision blockchain network; S160: querying and analyzing the information stored on the supervision blockchain network.

[0007] Optionally, the step S110 comprises: S1101: building a corresponding consortium blockchain network according to the scope of supervision and the involved institutions; S1102: granting different authorities to different institutions.

[0008] Optionally, the step S120 comprises: S1201, designing uplink information, the uplink information comprising one or a combination of a timestamp, user identity information, input data, generated content, and identification information of a generative AI model; S1202, designing a smart contract; and S1203, deploying the smart contract to the regulatory blockchain network.

[0009] Optionally, the step S1202 comprises: S12021, controlling the use permission of a user and limiting the use of a generative AI by an unauthorized user; S12022, obtaining transaction information and verifying the validity of the transaction information, the transaction information comprising one or a combination of generated text content, generation time, and input data information; and S12023, uploading the transaction information to a blockchain when the transaction information is valid.

[0010] Optionally, the step S130 comprises: S1301, receiving an authorization request sent by a user to the regulatory blockchain network; S1302, judging whether a user signature contained in the authorization request is correct based on the regulatory smart contract; S1303, in response to the regulatory smart contract judging that the user signature is correct, verifying the authorization request as a request sent by a correct user; S1304, in response to the authorization request sent by the correct user, obtaining input data of the user and judging whether the input data of the user is health data; S1305, in response to the data input by the user being health data, judging whether a state of the user is a normal state; and S1306, in response to the state of the user being the normal state, returning authorization information to the user.

[0011] Optionally, the step S140 comprises: S1401, in response to an authorized user using an AI model, generating text information corresponding to the input data of the user through the AI model based on the input data of the user.

[0012] Optionally, the step S150 further comprises: S1501, storing the generated text information, user information, identifier information of the AI model, and input data information in a blockchain network.

[0013] Optionally, the step S1501 comprises: binding the generated text information, user information, identifier information of the AI model, and input data information, and sending the bound data to the regulatory blockchain network for storage as one data.

[0014] Optionally, the step S160 comprises: S1601: querying the information stored on the supervision blockchain network; S1602: analyzing whether the generated text information is health information based on the query result; S1603: when the generated text information is non-health information, tracing the user information and model information bound to the generated text information; S1604: judging whether the non-health information is generated by the user or by the AI model based on the tracing result; S1605: generating analysis result prompt information based on the judgment result.

[0015] In another aspect, the present application provides a text generation type AI supervision system based on blockchain technology, comprising: a construction unit adapted to construct a supervision blockchain network; a deployment unit adapted to deploy a supervision smart contract on the supervision blockchain network; a judgment unit adapted to determine whether to authorize a user to use an AI model based on the supervision smart contract; a generation unit adapted to generate text information corresponding to the user using the AI model in response to the authorization; a storage unit adapted to store the text information and user information in the supervision blockchain network; a query and analysis unit adapted to query and analyze the information stored on the supervision blockchain network.

[0016] In yet another aspect, the present application provides a computer readable storage medium having one or more instructions stored therein, the computer instructions being used to make the computer execute the above-mentioned text generation type AI supervision method based on blockchain technology.

[0017] In still another aspect, the present application provides an electronic device, comprising: a memory and a processor; the memory has at least one program instruction stored therein; the processor loads and executes the at least one program instruction to realize the above-mentioned text generation type AI supervision method based on blockchain technology.

[0018] The beneficial effects of the present application are: the present application provides a text generation type AI supervision method based on blockchain technology, the method comprises the following steps: S110, constructing a supervision blockchain network; S120, deploying a supervision smart contract on the supervision blockchain network; S130, determining whether to authorize a user to use an AI model based on the supervision smart contract; S140, in response to the user being authorized to use the AI model, generating text information corresponding thereto; S150, storing the text information and user information into the supervision blockchain network; S160, querying and analyzing the information stored on the supervision blockchain network. By constructing a blockchain network, designing and deploying a supervision special-purpose smart contract, storing user information, input data and AI output data onto the blockchain, and allowing the supervision department to query and trace, the problem that the prior art cannot efficiently supervise the content generated by the text generation type AI is improved, the transparency and traceability of the generative AI are improved, thereby helping the government supervision department to better supervise and audit the use of the generative AI. BRIEF DESCRIPTION OF DRAWINGS

[0019] The present application will be further described below in conjunction with the drawings and examples.

[0020] Figure 1 is a flowchart of the text generation type AI supervision method based on the blockchain technology provided by the present application.

[0021] Figure 2 is a system schematic diagram of the text generation type AI supervision method based on the blockchain technology provided by the present application.

[0022] Figure 3 is a partial principle block diagram of the electronic device provided by the present application. DETAILED DESCRIPTION

[0023] Before discussing the example embodiments in more detail, it should be mentioned that some of the example embodiments are described as processes or methods depicted as flow charts. Although the processes are described in a particular sequential order, many of the processes can be performed concurrently, in parallel, or simultaneously. In addition, the order of the processes can be re-arranged. The processes can terminate when their operations are completed, but can also have additional steps not included in the figure. The processes can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0024] It should be understood that, although the terms "first", "second", etc. can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of the example embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] The application will now be described in detail with reference to the drawings. The diagram is a simplified schematic diagram, and only illustrates the basic structure of the application in a schematic manner, and thus only shows the components related to the application.

[0026] For ease of understanding, the professional terms appearing below are explained:

[0027] Generative AI: Generative AI refers to an AI that learns the components of an object from data through various machine learning (ML) methods, and then generates brand new, completely original content (such as text, pictures, and videos). The generated content is very similar to the training data, but it is not simply a copy of the learning data.

[0028] OpenID Connect: OpenID Connect is a simple identity protocol and open standard built on the OAuth 2.0 protocol. It allows client applications to rely on the authentication performed by the OpenID Connect provider to verify the identity of a user. OpenID Connect uses OAuth 2.0 for authentication and authorization, and then builds an identity for uniquely identifying a user. The client can also obtain basic profile information about the user from the OpenID Connect provider through interoperability and REST-like methods.

[0029] OAUTH: (Open Authorization) provides a secure, open and easy standard for user resource authorization. Unlike previous authorization methods, OAUTH authorization does not allow third parties to access user account information (such as username and password), i.e. third parties do not need to use the user's username and password to apply for authorization of the user's resources, so OAUTH is secure.

[0030] Embodiment 1

[0031] Please refer to Figure 1 The text generative AI supervision method based on blockchain technology described in the embodiments of the present application is shown in the flowchart.

[0032] As an example, the method comprises: S110: constructing a regulatory blockchain network; S120: deploying a regulatory smart contract on the regulatory blockchain network; S130: determining whether to authorize a user to use an AI model based on the regulatory smart contract; S140: generating text information corresponding thereto in response to authorizing the user to use the AI model; S150: storing the text information and user information into the regulatory blockchain network; S160: querying and analyzing the information stored on the regulatory blockchain network.

[0033] Optionally, the step S110 comprises:

[0034] S1101: building a corresponding consortium blockchain network according to the scope of regulation and the involved institutions; S1102: granting different authorities to different institutions. Wherein, granting different authorities to different institutions can comprise: when an institution is a supervisor, it can be granted a read permission, and when an institution is an owner of an AI generated model, it can be granted a read and write permission.

[0035] Optionally, the step S120 comprises:

[0036] S1201: designing on-chain information, the on-chain information comprising one or a combination of a timestamp, user identity information, input data, generated content, and identification information of a generative AI model; S1202: designing a smart contract; S1203: deploying the smart contract into the regulatory blockchain network. Wherein, the step S1202 comprises: S12021: controlling the use permission of a user and limiting the use of a generative AI by an unauthorized user; S12022: obtaining transaction information and verifying the validity of the transaction information, the transaction information comprising one or a combination of generated text content, generation time, and input data information; S12023: uploading the transaction information to the blockchain when the transaction information is valid.

[0037] Optionally, the generated content is text information generated by the AI model based on the input data of the user. In the process of designing the on-chain information, since there may be no generated content information generated at present, the generated content information can be designed with an empty character. When the generated content information is generated later, the generated text information can be stored in the pre-empty string.

[0038] Optionally, the control of the use permission of the user and the restriction of the unauthorized use of the generative AI include: determining the use permission according to the information of the user; or determining the use permission of the user based on the signature of the user. Specifically, a user signature is uploaded when the blockchain is called. The user signature can be encrypted and decrypted through the key mode, and finally the use permission of the user is determined based on the decrypted user signature. Specifically, identity authentication can be realized by using standards such as OAuth and OpenID Connect. Since the standards such as OAuth and OpenID Connect are very mature in the prior art, they will not be described here.

[0039] Optionally, verifying the validity of the transaction information can include: determining whether it is valid through the user signature, judging whether the signature of the user is consistent with the pre-stored data, if it is consistent, the transaction information is determined to be valid, and if it is not consistent, it is invalid.

[0040] Optionally, the step S130 includes: S1301: receiving an authorization request sent by a user to the regulatory blockchain network; S1302: determining whether the user signature contained in the authorization request is correct based on the regulatory smart contract; S1303: in response to the regulatory smart contract determining that the user signature is correct, verifying the authorization request as a request sent by a correct user; S1304: in response to the authorization request sent by the correct user, obtaining the input data of the user and determining whether the input data of the user is health data; S1305: in response to the data input by the user being health data, determining whether the state of the user is normal; S1306: in response to the state of the user being normal, returning authorization information to the user.

[0041] Optionally, step S1304: in response to the authorization request sent by the correct user, obtaining the input data of the user and determining whether the input data of the user is health data includes: determining whether the data input by the user is health data based on the keyword in the data input by the user. Specifically, before the user expects to use the AI model, the user first needs to send an authorization request to the blockchain network, and the authorization request carries the information of the user and the keyword input by the user. Based on the pre-deployed regulatory smart contract, it is first determined whether the user signature is correct to verify whether it is a request sent by a correct user, then the input content of the user is judged to verify whether the user tries to input illegal content to the model, and finally the state of the user is determined to verify whether the user is disabled. If it is a correct user who inputs legal content and the state of the user is normal, the authorization information is returned. That is, the above three discrimination conditions, if one of the discrimination conditions of the user is not met, the authorization information will not be returned to the user.

[0042] Optionally, the step S140 comprises: S1401: in response to the authorized user using the AI model, generating text information corresponding to the input data of the user based on the AI model.

[0043] Optionally, the step S150 further comprises: S1501: storing the generated text information, user information, identifier information of the AI model, and input data information in the blockchain network. Wherein, the step S1501 comprises: binding the generated text information, user information, identifier information of the AI model, and input data information, and sending the bound data to the supervision blockchain network for storage. In this way, it is convenient for subsequent tracing based on the generated text information.

[0044] Optionally, the step S160 comprises: S1601: querying the information stored on the supervision blockchain network; S1602: analyzing whether the generated text information is health information based on the query result; S1603: when the generated text information is non-health information, tracing the user information and model information bound to the generated text information; S1604: judging whether the non-health information is generated by the user or the AI model based on the tracing result; S1605: generating analysis result prompt information based on the judgment result. Specifically, when a user generates text using generative AI, the transaction information of the user will be written into the blockchain, and the transaction information should include the generated text content, generation time, input data, and the like. The supervision department can access the data on the blockchain through a query interface to query the text generated by a user within a certain period of time, and further analyze whether the text content is legal. If it is illegal, trace the user and the model, judge and determine whether the illegal information is caused by the user or the model, if it is caused by the user, limit the user from calling again, or even investigate legal responsibility; if it is caused by the model, limit the model from being called again.

[0045] Through the above embodiments of the present application, the present application improves the problem that the prior art cannot efficiently supervise the content generated by the text generative AI by constructing a blockchain network, designing and deploying a supervision special smart contract, storing user information, input data, and AI output data on the blockchain, and allowing the supervision department to query and trace, improving the transparency and traceability of the generative AI, thereby helping the government supervision department to better supervise and audit the use of the generative AI.

[0046] Embodiment 2

[0047] Please refer to 2, the present embodiment provides a text generative AI supervision system based on blockchain technology.

[0048] As an example, the system comprises:

[0049] The construction unit 210 is suitable for constructing a regulatory blockchain network.

[0050] The deployment unit 220 is suitable for deploying a regulatory smart contract on the regulatory blockchain network.

[0051] The judgment unit 230 is suitable for determining whether to authorize a user to use an AI model based on the regulatory smart contract.

[0052] The generation unit 240 is suitable for generating text information corresponding to the user using the AI model in response to the authorization.

[0053] The storage unit 250 is suitable for storing the text information and user information in the regulatory blockchain network.

[0054] The query and analysis unit 260 is suitable for querying and analyzing the information stored in the regulatory blockchain network.

[0055] Embodiment 3

[0056] The embodiment of the present application also provides a computer readable storage medium, wherein one or more instructions are stored in the computer readable storage medium, and the temperature control program of the tea making machine is executed by a processor to implement the steps of the text generation type AI supervision method based on the blockchain technology as described above. Since the storage medium adopts all the technical solutions of the above-mentioned embodiments, it at least has all the beneficial effects brought by the technical solutions of the above-mentioned embodiments, which will not be repeated here.

[0057] Embodiment 4

[0058] Please refer to Figure 3 The embodiment of the present application also provides an electronic device, which comprises a memory 302 and a processor 301; the memory 302 stores at least one program instruction; and the processor 301 loads and executes the at least one program instruction to implement the text generation type AI supervision method based on the blockchain technology as provided in the embodiment 1.

[0059] The memory 302 and the processor 301 are connected using a bus arrangement which can include any number of interconnecting buses and bridges needed to support their operation, and can be connected together via a bus or buses 305. The bus 305 can also connect various other circuits of the device, voltage regulators, power management circuits, and the like, well known in the art, thus, not all of these are shown or described in greater detail herein. A bus interface provides an interface between the bus 305 and a transceiver. The transceiver can be a single device or a plurality of devices, such as a plurality of receivers and transmitters, which are arranged to transmit and receive signals over a transmission medium. Data processed by the processor 301 is transmitted over a wireless medium via an antenna, and further, the antenna receives data and transmits the data to the processor 301.

[0060] The processor 301 is responsible for managing the bus and general processing, which can also include the functions of an operating system, application programs and the like. The memory 302 can be used for storing data used by the processor 301 when executing programs.

[0061] The above description is only an embodiment of the present application, and the common knowledge of the specific structure and characteristics in the scheme is not described in detail. The ordinary skilled person in the art knows all the ordinary technical knowledge in the field of the application before the application date or the priority date, can know all the prior art in the field, and has the ability to apply conventional experimental means before that date. The ordinary skilled person in the art can improve and implement the present scheme based on the disclosure given in this application, and some typical known structures or known methods should not be an obstacle for the ordinary skilled person in the art to implement the present application. It should be noted that for those skilled in the art, without departing from the structure of the present application, a number of modifications and improvements can be made, which should be considered as the protection scope of the present application, and these will not affect the effect and practicality of the present application. The scope of protection claimed in this application should be subject to the content of its claims, and the specific embodiments described in the specification can be used to explain the content of the claims.

Claims

1. A text-generating AI-based regulatory method based on blockchain technology, characterized in that, The method includes: S110: Building a regulatory blockchain network; S120: Deploy regulatory smart contracts on the regulatory blockchain network; S130: Determine whether to authorize the user to use the AI ​​model based on the aforementioned regulatory smart contract; S140: In response to an authorized user using the AI ​​model, generate corresponding text information; S150: Store the text information and user information in the regulatory blockchain network; S160: Query and analyze the information stored on the regulatory blockchain network; Step S120 includes: S1201: Design on-chain information, which includes one or a combination of timestamps, user identity information, input data, generated content, and identifier information of generative AI models; S1202: Design smart contracts; S1203: Deploy the smart contract to the regulatory blockchain network; Step S1202 includes: S12021: Control user access permissions and restrict unauthorized users from using generative AI; S12022: Obtain transaction information and verify the validity of the transaction information, wherein the transaction information includes one or a combination of generated text content, generation time, and input data information; S12023: When the transaction information is valid, upload the transaction information to the blockchain.

2. The text-generating AI supervision method based on blockchain technology according to claim 1, characterized in that, Step S110 includes: S1101: Establish a corresponding consortium blockchain network based on the scope of regulation and the institutions involved; S1102: Grant different permissions to different organizations.

3. The text-generating AI supervision method based on blockchain technology according to claim 1, characterized in that, Step S130 includes: S1301: Receive the authorization request sent by the user to the regulatory blockchain network; S1302: Determine whether the user signature contained in the authorization request is correct based on the regulatory smart contract; S1303: In response to the supervisory smart contract determining that the user signature is correct, the authorization request is verified as a request sent by the correct user; S1304: In response to a valid authorization request from a user, obtain the user's input data and determine whether the user's input data is healthy data; S1305: In response to user input data being health data, determine whether the user's status is normal; S1306: In response to the user's status being normal, return authorization information to the user.

4. The text-generating AI supervision method based on blockchain technology according to claim 1, characterized in that, Step S140 includes: S1401: In response to an authorized user using the AI ​​model, generate corresponding text information through the AI ​​model based on the user's input data.

5. The text-generating AI supervision method based on blockchain technology according to claim 1, characterized in that, Step S150 further includes: S1501: Store the generated text information, user information, AI model identifier information, and input data information in the blockchain network.

6. The text-generating AI supervision method based on blockchain technology according to claim 5, characterized in that, Step S1501 includes: The generated text information, user information, AI model identifier information, and input data information are bound together, and the bound data is sent as a single data set to the regulatory blockchain network for storage.

7. The text-generating AI supervision method based on blockchain technology according to claim 1, characterized in that, Step S160 includes: S1601: Query the information stored on the regulatory blockchain network; S1602: Analyze whether the generated text information based on the query results is health information; S1603: When the generated text information is unhealthy information, trace the user information and model information bound to the generated text information; S1604: Based on the traceability results, determine whether the non-health information was generated by the user or by an AI model; S1605: Generate analysis result prompt information based on the judgment result.

8. A text-generating AI-based monitoring system based on blockchain technology, characterized in that, The system includes: Building blocks, suitable for building regulatory blockchain networks; A deployment unit, adapted to deploy regulatory smart contracts on the regulatory blockchain network, includes: designing on-chain information, wherein the on-chain information includes one or a combination of timestamps, user identity information, input data, generated content, and identifier information of the generative AI model; designing a smart contract; and deploying the smart contract to the regulatory blockchain network; wherein the design of the smart contract includes: controlling user access permissions and restricting unauthorized users from using the generative AI; acquiring transaction information and verifying the validity of the transaction information, wherein the transaction information includes one or a combination of generated text content, generation time, and input data information; and uploading the transaction information to the blockchain when the transaction information is valid. The judgment unit is adapted to determine whether to authorize a user to use the AI ​​model based on the aforementioned regulatory smart contract; The generation unit is suitable for generating corresponding text information in response to authorized users using AI models; A storage unit, suitable for storing the text information and user information in the regulatory blockchain network; The query and analysis unit is suitable for querying and analyzing information stored on the regulatory blockchain network.

Images