Welcome to the China AI Bulletin
Fortnightly coverage of AI development, governance, and safety in China
No one working on AI governance can afford to underestimate the importance of China, a development—and regulatory—superpower.
But understanding China’s AI ecosystem is challenging: sources are in Chinese, regulation is fragmented across agencies, and the volume of technical, policy and standards information is almost impossible for any one actor to grasp at once.
The consequence is that research and discussions on international coordination frequently take place among actors with an incomplete or outdated picture of what people in China are actually doing.
The China AI Bulletin exists to close that gap. We monitor primary Chinese sources—government filings, standards publications, lab releases, and industry news—and provide broad coverage of AI development, governance, and safety while distilling what matters for people working on international AI coordination and competition.
Each briefing starts with with an executive summary, so you can get a high-level understanding of the field in 60 seconds or less, before diving into details that are crucial for substantive engagement.
We cover five primary areas:
Frontier Lab Developments: who released what, what the most important releases are, and why they matter in context.
Technical AI Safety Publications: what Chinese researchers are publishing on alignment, robustness, evaluations, and other safety-relevant topics.
International AI Governance: where China is engaging in bilateral summits and cross-border coordination.
Domestic AI Governance: regulatory moves from centralized bodies, key standards publications, and other key domestic developments.
Export Controls & Economic Policy: chips, trade restrictions, and the industrial policy context.
If you’re working on AI governance and need to stay current on China… welcome to the China AI Bulletin.
The China AI Bulletin is maintained by the Safe AI Forum (SAIF), a US 501(c)3 facilitating international cooperation on extreme AI risks. Views expressed represent individual authors’ perspectives, not official SAIF positions.
