China Interim Measures for Generative AI Services
China's binding rules for generative AI services offered to the public in China, including through APIs. Providers answer for lawful training data and content, must label generated images and video, handle complaints, and remove illegal output. Services with public-opinion or social-mobilisation capacity need a security assessment and an algorithm filing. Non-public R&D is excluded.
At a glance
In force
- Jurisdiction
- China
- Type
- Regulation
- Official title
- 生成式人工智能服务管理暂行办法 (Interim Measures for the Management of Generative Artificial Intelligence Services), CAC and six other departments, Order No. 15
What this means for you
If you build AI
- If your generative AI service is offered to the public in China, including through an API, you count as a provider and answer for training data, content controls and labelling.
- Services with 'public opinion attributes or social mobilisation capacity' need a security assessment and an algorithm filing before launch.
- Research and in-house use that is not offered to the public in China falls outside these Measures.
If you deploy AI
- Providers must stop generating and transmitting illegal content, fix the model, report it, and act against users who misuse the service.
- Generated images and video must be labelled under the deep synthesis rules; the 2025 labelling measures add detail.
If you use AI
- Providers in China must offer easy complaint and reporting channels and publish how they handle them.
- You can ask a provider to access, copy, correct or delete your personal information.
Key dates
- Aug 15, 2023PastInterim Measures take effectSource
Status history
In the Drops
No Drop has covered this law yet.