An ai copyright gateway implementation method and system based on timestamp public key signature

Abstract

The application discloses an AI copyright gateway implementation method and system based on a timestamp public key signature, and relates to the fields of information security and artificial intelligence. In view of the problems of difficult right certification and data leakage risk of generative artificial intelligence in generating digital content, the application proposes to deploy a gateway in front of an AI server, which is used for proxy forwarding of user requests and generated content responses. The gateway extracts a cryptographic hash digest of the AI generated content in real time, combines an authoritative timestamp and a digital certificate, uses a public key system to perform real-time digital signature on the "user identifier, content digest and timestamp", and generates a trusted copyright certificate. Optionally, the signature is further embedded in the content itself or stored on the chain as a hidden watermark. The application realizes automatic trusted right certification and tamper-proofing of AIGC content, effectively prevents internal leakage of core data through physical isolation and digest evidence leaving mechanism, and has the advantages of flexible deployment, high performance and meeting forward-looking cryptographic compliance.

Description

Technical Field

[0001] This invention relates to the fields of information security, cryptography, and artificial intelligence, and in particular to a gateway implementation method and system for real-time digital copyright protection, trusted evidence retention, and leakage prevention of content generated by generative artificial intelligence (AIGC). Background Technology

[0002] With the rapid development of generative artificial intelligence (AI) technology, text, images, audio, video, and code generated by large AI models are gradually becoming important core assets and work outcomes for users. However, in the current AI application ecosystem, the digital intellectual property protection of AIGC content faces severe challenges: 1. Difficulty in establishing rights and providing evidence: Once an intellectual property dispute arises, due to the lack of neutral and credible records of the generation process, it is extremely difficult for users to prove that a certain piece of content was legally generated by AI at a specific time. 2. Data leakage risk: Existing AI servers typically process and store user prompts and generated content in plaintext. System administrators, AI developers, or operations and maintenance personnel may have unauthorized access, leading to the theft or misuse of users' design ideas and core secrets. 3. Lack of automated copyright protection mechanism: Currently, the industry lacks a transparent, seamless automated copyright protection infrastructure that can connect with authoritative third parties (such as CA institutions and copyright bureaus). 4. While there are indeed some digital signature or timestamp signature servers on the market, they are mostly products that provide basic cryptographic calculation API interface services. Applications / servers need to actively request signature services through complex data encapsulation, which cannot achieve automatic copyright protection. Therefore, there is an urgent need for a copyright automation gateway solution that can solve the above problems. Summary of the Invention

[0003] The purpose of this invention is to provide an AI copyright gateway implementation method and system based on timestamp public key signature, in order to solve the problems of difficulty in confirming the rights of AIGC content and easy leakage in the prior art. The core technical solution of this invention is as follows: An AI copyright gateway system based on timestamp public key signature, characterized in that it includes: A. Network Proxy and Interception Module: The gateway is deployed in front of the AI ​​large model server. This module is used to receive AI generation requests sent by clients and forward the requests to the backend AI server; at the same time, it receives the generated content response returned by the AI ​​server; the working mode supports reverse proxy mode and transparent gateway mode. B. Cryptographic Processing and Signature Module: Supports international standard cryptographic algorithms (such as RSA, ECDSA), national commercial cryptographic algorithms (such as SM2, SM3, SM4), and quantum-resistant cryptographic algorithms (PQC); used to extract features of generated AI content, and combine digital certificates issued by authoritative CA institutions and public key timestamp services to perform real-time digital signatures on the generated content; C. Storage Module: Responsible for securely storing timestamped signature information, user identification, and content summary. An AI copyright gateway implementation method based on timestamp public key signature, characterized by comprising the following steps: S1. Access Request and Authentication: The gateway receives AI service requests from user terminals, parses the request protocol, and verifies the legitimacy of the user's identity; S2. Request Forwarding and Content Generation: The gateway securely forwards legitimate requests to the backend AI server; the AI ​​server processes the request and generates the corresponding AIGC content, which is then sent back to the gateway. S3. Real-time feature extraction and privacy protection: After receiving the generated content, the gateway extracts the cryptographic hash digest of the content; in privacy protection mode, the gateway only retains the hash digest and immediately releases or forwards the original plaintext content to prevent leakage on the gateway side; S4. Authoritative Timestamp and Digital Signature: The gateway calls the authoritative timestamp server and internal cryptographic module to digitally sign the "user identifier + digest + timestamp" using the PKI public key system, generating a trusted copyright certificate; S5. Secure Storage: The gateway securely stores signature or credential information in a database or file system; S6. Content returned to user: The gateway replies the final content, with an appended digital signature, to the user terminal. Optionally, the system further includes a watermarking module, which dynamically embeds the cryptographic signature or hash digest of the generated content into the generated image, video, or audio data stream in the form of an invisible digital watermark, thereby achieving a deep physical binding between the encryption features and the media file. Optionally, the system also includes an on-chain module responsible for sending timestamped signature information, user identification, and content summary to the judicial consortium blockchain node. Optionally, the method is characterized in that it does not have an interception module in the form of a reverse proxy or transparent proxy, but provides content signature invocation service to the user terminal and AI server in the form of an API interface. Optionally, the method is characterized by not calling an external authoritative timestamp server, but instead calling the time synchronization module and cryptographic module equipped on the gateway itself. The beneficial effects of this invention are as follows: 1. Automatic and Reliable Rights Confirmation: The content generated by each AI request from a user can be automatically and transparently recorded as reliable evidence, without the need for manual intervention and a huge amount of work each time; this greatly reduces the difficulty of providing evidence in subsequent intellectual property disputes. 2. Cut off internal leakage channels: Through the physical isolation and hash digest evidence retention mechanism of the gateway, there is no need to store sensitive results in plaintext on the AI ​​server for a long time, which effectively prevents illegal theft by system administrators or maintenance personnel; or even if the plaintext content is illegally stolen, its true copyright ownership can be easily determined later. 3. Flexible deployment and high performance: The system supports multiple deployment methods such as reverse proxy and transparent connection, requiring zero modification to business systems; combined with underlying hardware acceleration technology, it can perfectly adapt to high-concurrency, high-bandwidth AI large-scale model application scenarios. 4. Forward-looking security and compliance: Fully supports domestically developed commercial cryptographic algorithms and quantum-resistant algorithms, complies with relevant national data security and cryptographic application compliance requirements, and has long-term defense capabilities for future cybersecurity environments. Attached Figure Description Figure 1 This is a schematic diagram illustrating the implementation process of an AI copyright gateway based on timestamp public key signature according to the present invention. Figure 2 This is a schematic diagram of the method flow of Embodiment 2 of the present invention. Detailed Implementation The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments. Example 1: This embodiment describes the basic workflow of the gateway in reverse proxy or transparent gateway mode. The gateway is deployed in front of the AI ​​server through a network proxy and interception module. When a user initiates a request, the gateway executes S1 to request access and authenticate, parsing the protocol and verifying the identity. Then, it executes S2 to forward the request to the AI ​​server and receive the generated AIGC content. After obtaining the content, it proceeds to S3 for real-time feature extraction and privacy protection, where the gateway extracts the cryptographic hash digest and immediately releases the original plaintext content in privacy protection mode. Next, the cryptographic processing and signature module executes S4 to digitally sign the "user identifier + digest + timestamp" combination with the digital certificate and timestamp. The generated credential is securely stored in the local database via S5, and finally, the digitally signed content is returned to the user via S6. Example 2: This embodiment, based on Embodiment 1, adds a high-strength physical binding and a third-party consensus mechanism, making it suitable for multimedia generation scenarios with extremely high copyright protection requirements (such as AI video and image generation). In this embodiment, after executing the S4 certificate generation, the following steps are inserted: The gateway calls the watermarking module to dynamically embed the calculated signature or certificate information as an invisible watermark into the AIGC media content itself, achieving deep physical binding. After the content is returned to the user, the following steps are added: The gateway calls the on-chain module to send the calculated signature or certificate information (signature information with a timestamp, user identity identifier, and content summary) to a third-party trusted evidence storage institution or judicial consortium blockchain node, thereby forming an immutable distributed evidence chain.

Claims

1. An AI copyright gateway system based on timestamped public key signatures, characterized in that, include: A. Network Proxy and Interception Module: The gateway is deployed in front of the AI ​​large model server. This module is used to receive AI generation requests sent by clients and forward the requests to the backend AI server; at the same time, it receives the generated content response returned by the AI ​​server; the working mode supports reverse proxy mode and transparent gateway mode. B. Cryptographic Processing and Signature Module: Used to extract the features of the generated AI content, combine the digital certificate issued by an authoritative CA institution and the public key timestamp service, and use the public key signature algorithm to digitally sign the generated content in real time. C. Storage Module: Responsible for securely storing timestamped signature information, user identification, and content summary.

2. A method for implementing an AI copyright gateway based on timestamp public key signature, characterized in that, Includes the following steps: S1. Access Request and Authentication: The gateway receives AI service requests from user terminals, parses the request protocol, and verifies the legitimacy of the user's identity; S2. Request Forwarding and Content Generation: The gateway securely forwards legitimate requests to the backend AI server; the AI ​​server processes the request and generates the corresponding AIGC content, which is then sent back to the gateway. S3. Real-time feature extraction and privacy protection: After receiving the generated content, the gateway extracts the cryptographic hash digest of the content; in privacy protection mode, the gateway only retains the hash digest and immediately releases or forwards the original plaintext content to prevent leakage on the gateway side; S4. Authoritative timestamp and digital signature generation: The gateway calls the authoritative timestamp server and internal cryptographic module to digitally sign the "user identifier + digest + timestamp" using the PKI public key system to generate a trusted copyright certificate; S5. Secure Storage: The gateway securely stores signature or credential information in a database or file system; S6. Content returned to user: The gateway replies the final content, with an appended digital signature, to the user terminal.

3. The system according to claim 1, further comprising: Watermarking module: Dynamically embeds cryptographic signatures or hash digests of generated content into the generated image, video, or audio data stream as an invisible digital watermark, achieving deep physical binding between cryptographic features and media files.

4. The system according to claim 1, further comprising: On-chain module: responsible for sending the timestamped signature information, user identity identifier, and content summary to the judicial consortium blockchain node.

5. The system according to claim 1, characterized in that, Instead of using reverse proxy or transparent proxy to have an interception module, it provides content signing and invocation services to users and AI servers via API interfaces.

6. The method according to claim 2, characterized in that, Instead of calling an external authoritative timestamp server, it calls the high-precision time synchronization module or basic system time equipped on the gateway itself for signature calculation.

Images