China's NDRC says it is accelerating rules covering AI from development to use, citing loss-of-control risks
At a briefing, NDRC official Wang Ruomeng said the rules will span data to computing resources, require developers to disclose safety risks, incidents and vulnerabilities, and add sector sandboxes; no draft text has been published.
Why it matters
China already regulates generative AI services and recommendation algorithms piece by piece; a single framework covering the whole chain from data and compute to deployment would be broader, and it is expected to apply to DeepSeek, Moonshot AI and Alibaba. The NDRC also said it will monitor AI's effect on employment and provide a safety net for displaced workers. The details so far come from Korean press reports of the briefing. Watch for a published draft and what the disclosure duty requires of labs.
Key facts
- Announced at an NDRC briefing on November 3
- Speaker: Wang Ruomeng, NDRC innovation and high-tech department
- Scope: from data to computing resources
- Developers to disclose safety risks, incidents and vulnerabilities
- Sector regulatory sandboxes planned
- AI job-displacement monitoring planned