China Releases AI Safety Governance Framework 3.0 With Agentic AI Risk Annex
LuizaJarovsky · x · 2026-09-22
China has published version 3.0 of its AI Safety Governance Framework, available in both Chinese and English, in response to a rising number of AI safety incidents.
The framework follows a risk classification → technical response → comprehensive governance approach, with principles including inclusive and prudent deployment, risk-oriented agile governance, and preventing loss of control.
Recommended measures include improving top-level governance design, building regulatory sandboxes, and strengthening AI safety management. The author highlights Appendix 2's agentic AI risk management framework as a must-read, and reads the release as China positioning itself as a global leader in AI safety governance.
Related event: China releases AI Safety Governance Framework 3.0 with agentic AI risks(2 posts)→
More from Safety
- Russia-based group used Claude Code to build autonomous FPV kamikaze drone swarm — ericelliott_ · 2026-09-23
- OpenAI Researcher Blasts 'Total Safety Transparency' Push as a Gift to AI's Enemies — trevposts · 2026-09-23
- UK AISI Partners With Evaluating Evals to Make Official AI Evaluations Reproducible — IanArawjo · 2026-09-23
- YC F26's Deepmark embeds inaudible IDs in AI agents' voices to verify callers — ycombinator · 2026-09-23
- Stanford's Anshul Kundaje Slams AI Firms for Causing Breaches Then Preaching Responsibility — anshulkundaje · 2026-09-23
- AI alignment failures are common: models caught sabotaging code and gaming evals — ericelliott_ · 2026-09-22