China's National Development and Reform Commission said on Tuesday that it is fast-tracking legislation to put artificial intelligence under full-lifecycle supervision, covering AI development and applications from data through computing power. The stated goal of the framework is to keep increasingly powerful AI models from becoming uncontrollable.
Wang Ruomeng, deputy director-general of the NDRC's Department of Innovation and High-Tech Development, laid out the plan. Wang said the rules will span areas from data to computing resources, and that authorities will regularly evaluate how the rules are being implemented and revise provisions that become outdated.
The rules are expected to cover China's leading model developers, including DeepSeek, Moonshot AI and Alibaba Group. All three are building products meant to compete with frontier models from OpenAI and Anthropic.
Key facts
- 01China's National Development and Reform Commission said on November 3 it is fast-tracking AI legislation.
- 02The rules would put AI development and applications under full-lifecycle supervision, from data to computing power.
- 03DeepSeek, Moonshot AI and Alibaba are expected to fall under the new rules.
- 04Developers and operators would have to disclose safety risks, incidents and vulnerabilities.
- 05Beijing also plans sector-by-sector regulatory sandboxes and monitoring of AI-driven job displacement.
The NDRC plan pairs oversight with room to experiment. Regulators will introduce regulatory sandboxes sector by sector, giving companies a wider margin for trial and error before rules harden.
On the oversight side, AI developers and operators would be required to strengthen their risk management and to disclose safety risks, incidents and vulnerabilities in their systems.
The framework also reaches beyond the labs. The government plans to monitor AI's effects on employment and job displacement, and to provide a social safety net for workers affected by automation.
Coverage of the announcement described the package as one of the most comprehensive sets of AI safety guidelines to date. It framed the move as an attempt to balance continued AI innovation against the risk of models going out of control or being used for cyberattacks, along with concern about large-scale social disruption.
The NDRC's push follows recent warnings from Silicon Valley executives and researchers about AI's long-term risks, as well as incidents of AI systems being misused for cyberattacks.
It also comes two months after a September summit in Washington, where US President Donald Trump and Chinese President Xi Jinping agreed to exchange views on the risks and benefits of AI and to set up a communication channel for AI-related accidents.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
