China tackles over 3,500 non-compliant AI products since April

China's cyberspace watchdogs have addressed issues with over 3,500 artificial intelligence products that were not compliant with relevant rules since April, including mini-programs, web applications and AI agents.

This follows a campaign launched in April targeting the abuse of AI technology in forms such as deepfake face-swapping and voice-cloning, which infringe on public interests, as well as the failure to properly label AI-generated content that misleads the public, according to the Cyberspace Administration of China (CAC) on Friday.

More than 960,000 items containing illegal or harmful content were removed from the internet, and over 3,700 related accounts were shut down during the period, the CAC said.

During this phase, the CAC instructed local cyberspace authorities to intensify their actions against non-compliant AI products and to cut off their marketing and traffic channels. It urged major websites and platforms to strengthen technical safeguards. Efforts were also made to accelerate the implementation of labeling regulations for AI-generated content.

In the next phase of the campaign, the CAC will focus on prominent issues such as AI-generated rumors and vulgar online content, build a technical monitoring system, and standardize sanction protocols to maintain a healthy online environment and steer AI development in a more positive direction.