Avoid data bias

Avoid data security issues such as data ownership

Based on artificial intelligence data security risk research, relying on existing data security management mechanisms and technical means

Avoid data security risks such as training data pollution and data intelligence theft

Regulate prominent issues such as excessive collection of artificial intelligence-related data, bias and discrimination, abuse of resources, and deep forgery

Formulate and refine relevant departmental regulations to address key AI data security risks in various fields.

Strengthen safety supervision and inspection of high-risk links such as artificial intelligence data collection, use, and sharing, and conduct regular safety audits

Standardize methods and procedures for emergency response, investigation and evidence collection of artificial intelligence data security incidents

Implement supervision and inspection through various online and offline methods to timely discover and prevent Security risks

Establish technical monitoring means and public supervision channels

Establish an artificial intelligence data security detection and evaluation platform

Develop data security detection and evaluation methods and indicator systems for artificial intelligence products, applications and services

Develop a security testing and evaluation tool set to improve the safety and maturity of artificial intelligence products through testing and verification.

Break through the core key technologies of artificial intelligence data security protection such as small sample learning, federated learning, and differential privacy.