What happened
At an Aug 26, 2026 State Council Information Office press conference, MIIT Vice Minister Xin Guobin confirmed China has formulated nearly 200 key AI standards and that multiple cities are now operating AI ethics-review pilot programs requiring review before AI development begins (in principle, before model pre-training), making China's approach the only major jurisdiction requiring pre-development ethics review as an operational, locally enforced requirement rather than a published framework.
Why it matters
This marks China's AI governance architecture moving from published policy to operational, locally-enforced institutional compliance — a structurally different model from the EU's ex-post risk classification or the US's largely voluntary approach, with implications for any multinational developing or deploying models with Chinese entities or subsidiaries.
Action needed
Multinational AI developers with China-based R&D operations or partnerships should monitor which cities have activated ethics-review pilots and confirm applicability before beginning new model development work in-scope.