China Launches Special Crackdown on Illegal, Improper AI Activities

China has launched a four-month nationwide special campaign to crack down on illegal and improper AI activities.

The campaign will proceed in two phases. Phase 1 will focus on concentrated checks of issues including the omission of large model registration, insufficient platform security and review capabilities, use of illegal or non-compliant training data, AI data contamination, inadequate labeling of synthetic content, AI misuse (cyber attacks, deepfakes, voice cloning), and the absence of security management for open-source models. Phase 2 will focus on rectifying illegal service provisions such as distortion of classical works using AI, mass production of digital trash, generation and dissemination of false information, impersonation of public figures, creation of content restoring deceased individuals, generation of harmful content targeting minors, online public opinion manipulation, lewd chat services, one-click nudity, and the operation of fake AI websites and applications.

This is seen as an extension of the Chinese authorities' efforts to strongly control issues like disinformation, deepfakes, and harmful content arising from the spread of AI technology.