Measures for Artificial Intelligence Science and Technology Ethics Review and Services (Trial)

Summary

Trial measures for ethics review of AI research and development that may pose ethics risks to human dignity, public order, life and health or the environment. Universities, research institutes, medical institutions and companies set up AI ethics committees or use AI ethics review and service centres. Reviews focus on human well-being, fairness, controllability and trustworthiness, transparency and explainability, traceability of responsibility, and privacy. An annexed list names three kinds of AI R&D that need expert re-review: human-machine fusion systems, systems able to mobilise public opinion, and highly autonomous decision systems where safety or health is at risk.

Language

No English version of this instrument was found. The English title and the summary are the register's own, written from the sources below; for the exact wording use the Chinese text.

Safety topics

Status notes

Issued by MIIT and nine other bodies with a notice signed on 20 Mar 2026, and published on the MIIT website on 2 Apr 2026; the MIIT page gives a drafting date (成文日期) of 30 Mar 2026, the gov.cn page 20 Mar 2026. The Measures have 6 chapters and 37 articles and take effect on the date of issuance (Art. 37). They build on the 2023 Measures for Science and Technology Ethics Review (Trial) (CN-04). Arts. 4-8 promote ethics standards, ethics risk monitoring and early warning, testing and assessment, certification and consulting services, and support for small and micro enterprises. Expert re-review follows the initial review and is organised by provincial authorities or the competent departments (Arts. 21-25); emergency reviews are generally completed within 72 hours (Art. 27). Under Art. 26, activities already subject to registration, filing or approval under the deep synthesis, algorithm recommendation or generative AI service rules, where meeting ethics requirements is an approval condition or a supervision item, need no expert re-review. Violations are handled under the Cybersecurity, Data Security, Personal Information Protection and Science and Technology Progress laws (Art. 32).

Sources

Open the official source https://www.gov.cn/zhengce/zhengceku/202604/content_7064646.htm

Secondary source: https://www.miit.gov.cn/jgsj/kjs/wjfb/art/2026/art_2995f16b28504ddcbb604e918eb15759.html

Local copy

Verification

History

1. 2026-10-14notes: The document is signed 20 Mar 2026 and was published online 2–3 Apr 2026. The MI → Issued by MIIT and nine other bodies with a notice signed on 20 Mar 2026, and pu — Rewritten from the stored text: chapters and articles, Arts. 4-8 services, the expert re-review procedure, Art. 26 with its condition, the 72-hour emergency review and Art. 32. The online publication date now rests on the MIIT page (2 Apr 2026). The unverified 2025 draft moved to verification.note.2026-10-14-refresh-translation-fixes · Claude Code (claude-opus-5-5)
2. 2026-10-14summary: AI-specific ethics review regime (6 chapters, 37 articles). It is run through in → Trial measures for ethics review of AI research and development that may pose et — Rewritten from the stored Chinese text. The annex calls the three categories activities that need expert re-review; 'high-risk' is not in the text. 'Accountability' is now 'traceability of responsibility' (责任可追溯). Scope, the committee duty (Art. 9) and service centres (Art. 11) are taken from the articles.2026-10-14-refresh-translation-fixes · Claude Code (claude-opus-5-5)
3. 2026-10-14english_version: – → {"kind": "none found"} — No English version in the stored copies, and none found in a targeted search on 14 Oct 2026.manual-2026-10-14 · Claude Code (claude-opus-5-5)

Related records

S&T Ethics Review Measures China · In force

Cybersecurity Law amendment (AI article) China · In force

Citation

Ministry of Industry and Information Technology, with NDRC, Ministry of Education, Ministry of Science and Technology, Ministry of Agriculture and Rural Affairs, National Health Commission, People's Bank of China, Cyberspace Administration of China, Chinese Academy of Sciences, and China Association for Science and Technology. Measures for Artificial Intelligence Science and Technology Ethics Review and Services (Trial) [工信部联科〔2026〕75号]. 2026-03-20. https://www.gov.cn/zhengce/zhengceku/202604/content_7064646.htm. Status: In force; verified 2026-10-14 (level A). Local copy archive/CN-15/2026-10-14/official.html retrieved 2026-10-14 09:45:36 UTC, SHA-256 191eb94e86e7d37329ce9f0558effda133718c54261764afc28b0b662b444be2. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record CN-15.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.