China's NDRC says it will speed up a comprehensive AI law and lifecycle safety standards

China's National Development and Reform Commission (NDRC) has set out how it intends to tighten the rules around artificial intelligence over the next five years. Speaking on November 3 at a press conference in the agency's series on the 15th Five-Year Plan (2026–2030), Wang Ruomeng, deputy director-general of the NDRC's Department of Innovation and High-Technology Development, said the government would balance development and security and build safety into AI research, deployment and governance.

What the planning agency outlined

According to Xinhua Finance and Gelonghui reports of the briefing, Wang said China will take stock of its existing AI governance experience and speed up work on foundational, comprehensive AI legislation. Alongside a general law, the government plans to strengthen rules on shared building blocks such as data, computing power and algorithms, as well as on key application scenarios.

Wang also described safety standards and ethical guidelines intended to cover the entire AI lifecycle. Because the technology changes quickly, Wang said, the rules would be evaluated periodically and provisions that no longer fit would be revised.

A softer approach to compliance

The second strand concerns how oversight is applied. Wang said supervision should match the speed and reach of AI, with reasonable transition periods and requirements that are transparent and predictable. Regulators are to issue compliance guidance for specific sectors and support assessment, consulting and certification services meant to lower compliance costs for companies. The official also described AI safety as a shared responsibility of government bodies, industry associations, companies and the public.

On employment, the NDRC said it would monitor how AI reshapes jobs and work to add new occupations, such as model trainers and data annotators, to China's national occupational classification.

What remains unclear

The remarks are a statement of direction, not a bill. No draft text, timetable or list of standards was published, and the reports do not say which body would lead the drafting. Bloomberg, as summarized by TechFlow, described the plan as full-lifecycle oversight aimed at keeping increasingly capable systems under control, and noted that it would reach major developers such as DeepSeek, Moonshot and Alibaba.

China already regulates specific uses of AI, including recommendation algorithms, deep synthesis and generative AI services, through separate sets of rules. A single comprehensive law would bring these strands under one framework, but how far it would go, and when a draft might be put to the national legislature, has not been announced.