China's Cyberspace Administration brought into force interim measures on generative artificial intelligence services, formalising a registration regime for providers offering products with public-facing impact.
The rules require pre-deployment security assessments, training-data governance, content labelling, and complaint-handling mechanisms, with carve-outs for purely internal enterprise use cases.
The framework is widely seen as a workable compromise compared with earlier draft language, with explicit endorsement of the development of "trustworthy" generative AI as a strategic industry.
Implementation remains uneven across provinces, but compliance work has become a substantial fixture for major Chinese AI labs, including Baidu, Alibaba, Tencent, and a wave of well-funded startups.