China's CPC Central Committee and State Council have issued an opinion on new quality productive forces that includes an "AI+" action and an AI risk early-warning system

Key takeaways

• The opinion is a Party and State Council policy text, not a named AI rule. It states no duty, deadline or penalty for private companies.
• Item (10) calls for the AI+ action and says to build monitoring, early-warning and emergency-response systems so AI is safe, reliable and controllable.
• Item (15) says to strengthen governance in AI and biology, alongside cross-department joint supervision.
• The text names no agency to write AI rules. Coordination of tracking and evaluation is assigned to the Central Financial and Economic Affairs Office and the NDRC.

Is the opinion binding, and who issued it?

It was issued jointly by the CPC Central Committee and the State Council and released by Xinhua on 9 October 2026, with the gov.cn page stamped 2026-10-09 18:16. The text opens by saying that developing new quality productive forces is an internal requirement of high-quality development, and that the opinion is put forward to speed that work up.

The text does not describe itself as a rule with penalties. Its final section tells regions and departments to work under the Party's centralized leadership, and it warns officials against using the new quality productive forces label to misuse policies and against blind investment that causes major losses. It does not set a date on which anything takes effect, and its implementation section is addressed to regions and departments.

What does item (10) say about AI+ and AI safety?

Item (10) is headed as comprehensively implementing the "AI+" action. It calls for AI to transform traditional industries and to speed up scenario applications in intelligent connected new-energy vehicles, AI phones and computers, and humanoid robots. It also calls for breakthroughs in basic theory and core technology, a stronger supply of compute, algorithms and data, and national AI industry pilot bases and high-value application scenarios, arranged by region and by sector (our translation).

On safety, the item ends with one sentence:

构建技术监测、风险预警、应急响应体系，确保人工智能安全、可靠、可控。

In our translation: build technology-monitoring, risk-early-warning and emergency-response systems so that AI is safe, reliable and controllable. The text does not say who builds these systems, what they cover, or by when.

Where else does AI appear in the opinion?

Item (15), on market supervision, includes a short line:

加强人工智能、生物等领域治理。

In our translation: strengthen governance in AI, biology and other fields. The same item calls for stronger cross-department joint supervision and a data-driven supervision approach, without naming AI instruments.

AI also appears in item (5), which says to use AI to lead a change in the scientific research paradigm. Two nearby items cover related infrastructure without mentioning AI by name: item (9) on data-property systems, data infrastructure and a national integrated computing network, and item (17) on an efficient, convenient and safe mechanism for cross-border data flows. None of these items sets a rule for AI developers or deployers.

What should policy teams and companies do now?

Treat the text as a direction signal and record it in your China monitoring file. The text announces no filing, registration or compliance step.

• Log items (10) and (15) with the issue date, and note that the text names no implementing measures, and watch for any implementing measures.
• Map where your products touch the areas item (10) names: intelligent connected new-energy vehicles, AI phones and computers, and humanoid robots. Separately, note item (17) on cross-border data flows.
• Keep existing obligations separate. This text does not replace or amend rules issued by the Cyberspace Administration of China, which we track in our CAC coverage.

What are the key dates and parties?

What we did not verify

What we opened: the full text of the opinion on gov.cn, in Chinese, on 2026-10-25 UTC. Translations on this page are ours.

What we did not open: any official English version, any implementing measure, and any commentary on how the AI+ action will be carried out.

What we refuse to claim: that the opinion creates binding AI duties for private companies, that any agency has been assigned to build the monitoring system, or what follow-on rules will contain.

Related briefings

• AI Regulation News hub
• Asia-Pacific AI regulation tracker
• China CAC, NDRC and MIIT opinions on AI agents
• TC260 AI Safety Governance Framework 3.0
• NFRA guidance on safe AI development in banking and insurance
• Supreme People's Court opinions on adjudicating AI disputes