AI data security management method and system based on model context protocol

Abstract

The invention discloses an AI data security management method and system based on a model context protocol. According to the method, the model state, the environment parameters and the user session information can be collected and analyzed in real time, and a dynamic encryption strategy and a permission token are generated. And each node uploads an encryption gradient through an MCP protocol and adds differential privacy noise, the federal learning coordinator aggregates parameters by using a secure multi-party computing technology to generate a global model, and synchronously updates a model version to all nodes through the MCP protocol. The method and the system can be applied to intelligent customer service, medical diagnosis and other scenes, and data security and compliance are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the cross - technical field of artificial intelligence and data security, and particularly relates to an AI data security management method and system based on a model context protocol. Background Art

[0002] With the rapid development of AI technology, data privacy protection, model secure transmission, and dynamic access control have become problems to be solved urgently.

[0003] Traditional AI systems usually adopt fixed encryption and access control rules, and cannot adapt to dynamic model contexts, such as real - time resource status, user permission changes, etc. In addition, centralized data storage and transmission are vulnerable to attacks, while distributed solutions such as federated learning can protect privacy but lack fine - grained context awareness capabilities. Existing communication protocols (such as HTTP / RPC) do not deeply integrate security mechanisms, and additional encryption and auditing modules need to be designed, resulting in low efficiency. Summary of the Invention

[0004] Based on this, the embodiments of the present application provide an AI data security management method and system based on a model context protocol. The present application dynamically senses the model context (such as model version, user session, hardware resources), realizes adaptive encryption, fine - grained access control, and privacy protection, and solves the problems of rigid security policies and high privacy leakage risks in traditional solutions.

[0005] In a first aspect, an AI data security management method based on a model context protocol is provided. The method includes:

[0006] Collecting the context information of the model in real - time; wherein, the context information at least includes model status, environmental parameters, and user session information;

[0007] In response to a security management request initiated by a client, parsing the collected context information to generate a dynamic security policy; wherein, the dynamic security policy at least includes the selection of an encryption algorithm and the generation of a dynamic permission token;

[0008] Performing data encryption operations and access control according to the generated dynamic security policy, and recording operation logs;

[0009] Securely aggregating model parameters among distributed nodes to generate a global model, synchronizing and updating the model version to all nodes through the MCP protocol, and notifying all nodes to update access permissions.

[0010] Optionally, collecting the context information of the model in real - time includes:

[0011] Continuously monitoring the weight version and gradient changes of the model; specifically including recording the weight update situation and gradient values after each model training;

[0012] Obtain the CPU and GPU load conditions of the current running environment in real time; specifically, regularly sample the usage rate, temperature, and memory occupancy parameters of the CPU and GPU.

[0013] Capture the user's permission level, geographical location, and session status; specifically, record the IP address, device type, login time when the user logs in, and the role and permission level of the user in the system.

[0014] Optionally, parse the collected context information to generate dynamic security policies, including:

[0015] Determine the encryption algorithm according to the model status and user session information, specifically including selecting different encryption algorithms according to the data type.

[0016] Generate a dynamic permission token by combining the user role and model version, and the dynamic permission token includes the user role, allowed operations, model version, and expiration time.

[0017] Optionally, according to the generated dynamic security policy, perform data encryption operations and access control, including:

[0018] Verify whether the permission token provided by the user complies with the current security policy; specifically, check whether the signature in the JWT is valid, whether the expiration time has passed, and whether the user role and allowed operations match the request.

[0019] If the verification passes, encrypt the data according to the encryption algorithm specified in the policy; specifically, call the encryption library to perform the encryption operation, and store or transmit the encrypted data.

[0020] Record all encryption, decryption, and access control operations to generate an operation log; specifically, record the time of the operation, user ID, operation type, data type, encryption algorithm, and operation result.

[0021] Optionally, the secure aggregation of model parameters between distributed nodes to generate a global model includes:

[0022] Each node uploads the local gradient through the MCP protocol, and adds differential privacy noise that meets the preset noise level to the gradient value.

[0023] Adopt the SMPC protocol to perform collaborative calculation of encrypted gradients between multiple nodes to ensure that data is not leaked during the aggregation process.

[0024] Apply the aggregated gradient to the parameter update of the global model, and broadcast the updated model parameters to all nodes.

[0025] Optionally, synchronously updating the model version to all nodes through the MCP protocol includes:

[0026] Sending the updated model version information to all distributed nodes through the MCP protocol; wherein, the message of the MCP protocol includes the model version number, the update timestamp, and the hash value of the model parameters;

[0027] Sending an update instruction to each node to instruct the node to regenerate a dynamic permission token matching the new model version and update the local access control policy.

[0028] In a second aspect, an AI data security management system based on a model context protocol is provided, and the system includes:

[0029] A context awareness module for collecting context information of the model in real time; wherein, the context information at least includes the model state, environmental parameters, and user session information;

[0030] An MCP protocol engine for parsing the collected context information to generate a dynamic security policy in response to a security management request initiated by a client; wherein, the dynamic security policy at least includes the selection of an encryption algorithm and the generation of a dynamic permission token;

[0031] A security control module for performing data encryption operations and access control according to the generated dynamic security policy and recording operation logs;

[0032] A federated learning coordinator for securely aggregating model parameters among distributed nodes to generate a global model, synchronously updating the model version to all nodes through the MCP protocol, and notifying all nodes to update access permissions.

[0033] In a third aspect, an electronic device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the AI data security management method according to any one of the first aspects above.

[0034] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the AI data security management method according to any one of the first aspects above.

[0035] In a fifth aspect, a computer program product is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the AI data security management method according to any one of the first aspects above.

[0036] The beneficial effects brought by the technical solutions provided in the embodiments of the present application at least include:

[0037] (1) Enhanced security: By means of dynamic encryption and fine-grained access control, the risk of data leakage is reduced;

[0038] (2) Efficiency optimization: The context-aware strategy reduces redundant encryption operations and communication overhead.

[0039] (3) Compliance: Meets data privacy regulations such as GDPR and HIPAA. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, other implementation drawings can be obtained based on the provided drawings without creative efforts.

[0041] Figure 1 It is a flowchart of the steps of the AI data security management method provided by the embodiment of the present application;

[0042] Figure 2 It is a system architecture diagram provided by the embodiment of the present application;

[0043] Figure 3 It is a flowchart of the generation of the dynamic security policy provided by the embodiment of the present application;

[0044] Figure 4 It is a schematic diagram of the secure aggregation process of federated learning provided by the embodiment of the present application;

[0045] Figure 5 It is a schematic diagram of an electronic device provided by the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] In order to make the purpose, technical solutions and advantages of the present application more clear, the following further details the present application in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application and are not used to limit the present application.

[0047] In the description of the present invention, the terms "include", "have" and any variations are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units that are clearly listed, but may also include other steps or units that are inherent to these processes, methods, products or devices, or steps or units added based on further optimization schemes conceived from the present invention.

[0048] It can be seen from the background technology that the following technical problems exist in the prior art:

[0049] Static security policy: Traditional AI systems adopt fixed encryption and access control rules and cannot adapt to dynamically changing model contexts (such as real-time resource status, user permission changes).

[0050] Risk of privacy leakage: Centralized data storage and transmission are vulnerable to attacks, and distributed solutions such as federated learning lack fine-grained context awareness capabilities.

[0051] Disconnection between protocol and security: Existing communication protocols (such as HTTP / RPC) do not deeply integrate security mechanisms, and additional encryption and auditing modules need to be designed, resulting in low efficiency.

[0052] This application proposes an AI data security management method and system based on the MCP protocol. By dynamically perceiving model contexts (such as model version, user session, hardware resources), it realizes adaptive encryption, fine-grained access control, and privacy protection, and solves the problems of rigid security policies and high risk of privacy leakage in traditional solutions. It is specifically used to achieve data privacy protection, secure model transmission, and dynamic access control during the distributed AI model training and inference processes.

[0053] Please refer to Figure 1 , which shows a flowchart of an AI data security management method based on the model context protocol provided by an embodiment of this application. The method may include the following steps:

[0054] S1, Collect the context information of the model in real time.

[0055] Among them, the context information includes at least model status, environmental parameters, and user session information.

[0056] Specifically, through the built-in monitoring module, the weight version and gradient changes of the model are obtained in real time. The specific implementation method includes setting hooks during the model training process to capture the weight update situation and gradient values after each training. These information will be stored in the local database or memory for subsequent processing.

[0057] Regularly sample the CPU and GPU load conditions of the current running environment. The specific implementation method includes using monitoring tools (such as psutil or nvidia-smi) to regularly obtain parameters such as CPU usage, GPU usage, temperature, and memory occupancy. These parameters will be recorded in the log file and updated in real time.

[0058] Capture the user's permission level, geographical location, and session status when the user logs in. The specific implementation method includes recording the IP address, device type, login time, and the user's role and permission level when the user logs in. These information will be stored in the user session management module for access control when needed.

[0059] S2. In response to a security management request initiated by a client, parse the collected context information to generate a dynamic security policy.

[0060] Among them, the dynamic security policy at least includes the selection of encryption algorithms and the generation of dynamic permission tokens.

[0061] Specifically, receive a security management request initiated by the client. This request contains the user ID, the type of operation requested (such as model inference, parameter update, etc.), and the target model version. First, verify the legality of the request, including user authentication and request format verification.

[0062] According to the collected context information, parse the model state, environmental parameters, and user session information. The specific implementation methods include reading the weight version and gradient information of the model from the local database or memory, obtaining environmental parameters from the log file, and obtaining the user permission level and session status from the user session management module.

[0063] Generate a dynamic security policy according to the parsed context information. The specific implementation methods include:

[0064] Encryption algorithm selection: Select different encryption algorithms according to the data type (such as medical data, financial data, etc.). For example, use the SM4 algorithm for medical data and the ChaCha20-Poly1305 algorithm for other data.

[0065] Dynamic permission token generation: Generate a dynamic permission token (such as JWT) by combining the user role and the model version. The generated JWT token contains the user role (such as "researcher"), the allowed operations (such as "model_inference"), the model version (such as "v2.3"), and the expiration timestamp.

[0066] S3. According to the generated dynamic security policy, perform data encryption operations and access control, and record operation logs.

[0067] Specifically, verify whether the permission token provided by the user complies with the current security policy. The specific implementation methods include checking whether the signature in the JWT is valid, whether the expiration time has passed, and whether the user role and the allowed operations match the request.

[0068] If the verification passes, encrypt the data according to the encryption algorithm specified in the policy. The specific implementation methods include calling an encryption library (such as pycryptodome or cryptography) to perform the encryption operation, and storing or transmitting the encrypted data.

[0069] Perform access control according to the dynamic security policy to ensure that only authorized users can access a specific version of the model.

[0070] Record all encryption, decryption, and access control operations to generate an operation log. The specific implementation methods include recording the operation time, user ID, operation type, data type, encryption algorithm, and operation result.

[0071] S4. Securely aggregate the model parameters among the distributed nodes to generate a global model, and synchronize and update the model version to all nodes through the MCP protocol. Meanwhile, notify all nodes to update the access permissions.

[0072] Specifically, each node uploads the local gradient through the MCP protocol and attaches differential privacy noise. The specific implementation methods include adding differential privacy noise that conforms to a preset noise level (such as ε = 0.1) to the gradient value to ensure the privacy protection of data during transmission.

[0073] The federated learning coordinator uses secure multi-party computation (SMPC) to securely aggregate the encrypted gradients. The specific implementation methods include using the SMPC protocol to perform collaborative calculations of encrypted gradients among multiple nodes to ensure that the data is not leaked during the aggregation process.

[0074] Update the global model according to the aggregation result. The specific implementation methods include applying the aggregated gradients to the parameter update of the global model and broadcasting the updated model parameters to all nodes.

[0075] Send the updated model version information to all distributed nodes through the MCP protocol. The specific implementation methods include including the model version number, update timestamp, and hash value of the model parameters in the message of the MCP protocol.

[0076] Notify all nodes to update the local access permissions according to the new model version. The specific implementation methods include sending an update instruction to each node, instructing the node to regenerate a dynamic permission token that matches the new model version, and updating the local access control policy.

[0077] Please refer to Figure 2 , which shows the architecture diagram of an AI data security management system based on the model context protocol provided by an embodiment of the present application. The system may include:

[0078] The system includes the following core modules:

[0079] MCP protocol engine: Responsible for context synchronization, dynamic generation of security policies, and communication management.

[0080] Context awareness module: Real-time collect the model state (weight version, gradient), environmental parameters (CPU / GPU load), and user session (permissions, geographical location).

[0081] Security control module: Integrate dynamic encryption, access control, and auditing functions.

[0082] Federated Learning Coordinator: Securely aggregate model parameters among distributed nodes (supporting differential privacy and secure multi-party computation).

[0083] Specifically, a context awareness module for real-time collection of context information of the model; wherein the context information includes at least model status, environmental parameters, and user session information;

[0084] The MCP protocol engine is used to respond to security management requests initiated by the client, parse the collected context information to generate dynamic security policies; wherein the dynamic security policies include at least the selection of encryption algorithms and the generation of dynamic permission tokens;

[0085] The security control module is used to perform data encryption operations and access control according to the generated dynamic security policies, and record operation logs;

[0086] The federated learning coordinator is used to securely aggregate model parameters among distributed nodes to generate a global model, synchronize the updated model version to all nodes through the MCP protocol, and notify all nodes to update access permissions. As Figure 3 The technical flow chart is given, where:

[0087] The client initiates a request, attaching context information (such as user ID, model version).

[0088] The MCP protocol engine generates security policies (encryption algorithms, access permissions) based on the context.

[0089] The security control module verifies permissions and performs encryption operations.

[0090] The federated learning coordinator aggregates encrypted parameters and updates the global model.

[0091] The audit module records all operation logs, supporting post-event traceability.

[0092] Specifically, the client sends a request to the MCP protocol engine, and the request attaches the context information of the model. This context information includes at least model status, environmental parameters, and user session information. For example, the client may request to access a specific version of the model for inference operations and provide user authentication information and model version number.

[0093] After receiving the request, the MCP protocol engine first parses the attached context information to determine the model version and user role. This step is the basis for generating dynamic security policies, ensuring that subsequent security policies can be adjusted according to real-time context information.

[0094] Based on the parsed context information, the MCP protocol engine generates dynamic security policies. This includes selecting appropriate encryption algorithms and generating dynamic permission tokens (such as JWT). For example, if the context information indicates that the data type is medical data, the MCP protocol engine may select the SM4 encryption algorithm; at the same time, according to the user's role and the type of operation requested, a JWT token containing the role, permitted operations, model version, and expiration time is generated.

[0095] The generated dynamic policy is sent to the security control module. The security control module is responsible for verifying the validity of the permission token and performing encryption operations according to the dynamic policy. This includes checking whether the signature of the JWT token is valid, whether the user role is authorized to perform the requested operation, and whether the model version matches.

[0096] The security control module returns the verification result to the MCP protocol engine. If the verification is successful, the request will be allowed to continue; if the verification fails, the request will be rejected, and the security control module will notify the MCP protocol engine of the reason for rejecting the request.

[0097] According to the verification result of the security control module, the MCP protocol engine sends a response to the client. If the request is allowed, the client can continue with model inference or other operations; if the request is rejected, the client will receive an error message indicating that the request failed the security verification.

[0098] In summary, the innovations of this application include:

[0099] Dynamic security policy generation: Dynamically adjust encryption and access rules based on real-time context (such as model version, network latency).

[0100] Lightweight federated security aggregation: Embed secure multi-party computation (SMPC) in the MCP protocol to achieve efficient privacy protection.

[0101] Fine-grained context binding: Bind user permissions to model version and hardware environment to prevent unauthorized operations.

[0102] The following gives another two specific embodiments:

[0103] Embodiment 1: Secure inference of an intelligent customer service system.

[0104] Scenario: The user queries medical information through a voice assistant.

[0105] Implementation steps:

[0106] The context-aware module captures the user identity (anonymous ID) and the sensitivity of the query content (medical data).

[0107] The MCP protocol engine triggers dynamic encryption (SM4) and restricts the model version to "Medical Special v1.2".

[0108] The security control module verifies the JWT token and only allows the invocation of the specified model version.

[0109] Before the inference result is returned, the audit module records the operation log and desensitizes sensitive fields.

[0110] Example 2: Distributed model training.

[0111] Scenario: Cross-hospital joint training of a tumor detection model.

[0112] Implementation steps:

[0113] Each hospital node uploads its local gradient through the MCP protocol, and the protocol automatically attaches differential privacy noise (ε = 0.1).

[0114] The federated learning coordinator uses SMPC to securely aggregate the gradients and generate a global model.

[0115] The MCP protocol synchronously updates the model version to v3.0 and notifies all nodes to update their access permissions.

[0116] As Figure 4 Figure shows the schematic diagram of the secure aggregation process of federated learning. Among them, each node uploads encrypted gradients through the MCP protocol and attaches differential privacy noise. The federated learning coordinator uses secure multi-party computation (SMPC) to aggregate the parameters and generate a global model. The updated model version is synchronized to all nodes through the MCP protocol.

[0117] Specifically, after each hospital node (such as Hospital Node A, Hospital Node B, Hospital Node C) completes local model training, it generates gradient information. These gradient information are first encrypted through the MCP protocol to ensure the security and privacy of the data during transmission.

[0118] Before uploading the encrypted gradients, each node attaches differential privacy noise. This step is to further enhance privacy protection by introducing randomness to prevent the leakage of sensitive information. For example, when Hospital Node B uploads gradients, it will add corresponding noise according to the preset privacy protection parameters (such as ε = 0.1).

[0119] The encrypted and noise-attached gradient information is sent to the federated learning coordinator. The coordinator is responsible for collecting the gradient information from each node and preparing for secure aggregation.

[0120] Before aggregation, the federated learning coordinator verifies the digital signature of the received gradients to ensure the integrity of the data and the authenticity of the source. This step helps to prevent malicious nodes from sending forged or tampered data.

[0121] The federated learning coordinator uses secure multi-party computation (SMPC) technology to securely aggregate the collected encrypted gradients. SMPC allows multiple nodes to jointly compute the result of a function without revealing their respective data. Here, it is to aggregate the gradients to generate a global model. This process ensures the privacy and security of the data.

[0122] The aggregated gradients are used to update the parameters of the global model. The federated learning coordinator encapsulates the updated model parameters into a new version of the global model (such as v3.0).

[0123] The updated global model version is synchronized to all nodes participating in federated learning through the MCP protocol. This step ensures that all nodes can obtain the latest model version in a timely manner, thus guaranteeing the consistency and accuracy of the model.

[0124] With the release of the new model version, each node needs to update its local access permissions to ensure that only authorized users and systems can access and use the new version of the model. This step is achieved through the MCP protocol, ensuring the timeliness and effectiveness of access control.

[0125] In one embodiment, an electronic device is provided. The electronic device may be a computer, and its internal structure diagram may be as Figure 5 shown. The electronic device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the device is used to provide computing and control capabilities. The memory of the device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to manage data for AI data security based on the model context protocol. The network interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements an AI data security management method based on the model context protocol.

[0126] Those skilled in the art can understand that the structure shown in Figure 5 is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0127] In one embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above-mentioned AI data security management method based on the model context protocol.

[0128] In one embodiment of the present application, a computer program product is provided, including computer programs / instructions. When the computer program is executed by a processor, it implements the steps of the above-mentioned AI data security management method based on the model context protocol.

[0129] For the computer-readable storage medium and computer program product provided in this embodiment, their implementation principles and technical effects are similar to those of the above method embodiments, and will not be elaborated here.

[0130] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned method embodiments. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in M forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Symchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0131] The technical features of the above-mentioned embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0132] The above-mentioned embodiments only represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the patent application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. An AI data security management method based on a model context protocol, characterized in that, The method includes: Collecting the context information of the model in real time; wherein, the context information at least includes the model state, environmental parameters, and user session information; In response to a security management request initiated by the client, parsing the collected context information to generate a dynamic security policy; wherein, the dynamic security policy at least includes the selection of an encryption algorithm and the generation of a dynamic permission token; Performing data encryption operations and access control according to the generated dynamic security policy, and recording operation logs; Securely aggregating model parameters among distributed nodes to generate a global model, synchronizing and updating the model version to all nodes through the MCP protocol, and notifying all nodes to update access permissions simultaneously.

2. The AI data security management method according to claim 1, wherein Collecting the context information of the model in real time, including: Continuously monitoring the weight version and gradient changes of the model; specifically including recording the weight update situation and gradient values after each model training; Obtaining the CPU and GPU load conditions of the current running environment in real time; specifically including regularly sampling the usage rate, temperature, and memory occupancy parameters of the CPU and GPU; Capturing the user's permission level, geographical location, and session status; specifically including recording the IP address, device type, login time when the user logs in, and the user's role and permission level in the system.

3. The AI data security management method according to claim 1, wherein, Parsing the collected context information to generate a dynamic security policy, including: Determining the encryption algorithm according to the model state and user session information, specifically including selecting different encryption algorithms according to the data type; Generating a dynamic permission token in combination with the user role and model version, and the dynamic permission token includes the user role, allowed operations, model version, and expiration time.

4. The AI data security management method according to claim 1, wherein, Performing data encryption operations and access control according to the generated dynamic security policy, including: Verifying whether the permission token provided by the user conforms to the current security policy; specifically including checking whether the signature in the JWT is valid, whether the expiration time has passed, and whether the user role and allowed operations match the request; If the verification passes, encrypting the data according to the encryption algorithm specified in the policy; specifically including calling the encryption library to perform the encryption operation, and storing or transmitting the encrypted data; Recording all encryption, decryption, and access control operations to generate operation logs; specifically including recording the operation time, user ID, operation type, data type, encryption algorithm, and operation result.

5. The AI data security management method according to claim 1, characterized in that, The securely aggregating model parameters among distributed nodes to generate a global model includes: Each node uploads the local gradient through the MCP protocol, and adds differential privacy noise that meets the preset noise level to the gradient value; Using the SMPC protocol to perform collaborative calculation of encrypted gradients among multiple nodes to ensure that data is not leaked during the aggregation process; Applying the aggregated gradient to update the parameters of the global model, and broadcasting the updated model parameters to all nodes.

6. The AI data security management method according to claim 1, wherein, The synchronizing and updating the model version to all nodes through the MCP protocol includes: Sending the updated model version information to all distributed nodes through the MCP protocol; wherein, the message of the MCP protocol includes the model version number, update timestamp, and hash value of the model parameters. Send an update instruction to each node, instructing the node to regenerate a dynamic permission token that matches the new model version and update the local access control policy.

7. An AI data security management system based on a model context protocol, characterized in that, The system includes: A context awareness module for real-time collection of context information of the model; wherein, the context information at least includes model status, environmental parameters, and user session information; An MCP protocol engine for responding to a security management request initiated by a client, parsing the collected context information to generate a dynamic security policy; wherein, the dynamic security policy at least includes the selection of an encryption algorithm and the generation of a dynamic permission token; A security control module for performing data encryption operations and access control according to the generated dynamic security policy and recording operation logs; A federated learning coordinator for securely aggregating model parameters among distributed nodes to generate a global model, synchronizing and updating the model version to all nodes through the MCP protocol, and notifying all nodes to update access permissions.

8. An electronic device, characterized in that, It includes a memory and a processor, and the memory stores a computer program, and when the computer program is executed by the processor, it implements the AI data security management method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and when the computer program is executed by a processor, it implements the AI data security management method according to any one of claims 1 to 6.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, it implements the AI data security management method according to any one of claims 1 to 6.

Images