
Beijing Institute of AI Safety and Governance (Beijing-AISI)
The Beijing Institute of AI Safety and Governance (Beijing-AISI), formally named in Chinese as 北京前瞻人工智能安全与治理研究院, is a research institute established under the oversight of Beijing's Municipal Commission of Economy and Information Technology. Founded in September 2024 as a laboratory and formally incorporated as an institute in May 2025, it is directed by Yi Zeng, a professor at the Chinese Academy of Sciences and a member of the UN High-Level Advisory Body on AI. The institute conducts fundamental research on AI safety, develops evaluation benchmarks such as ForesightSafety-Bench and PandaGuard (an LLM jailbreak research framework), formulates safety standards, and engages in international AI safety initiatives including the Paris AI Action Summit and the Singapore Conference on AI (SCAI 2025).
Funding Details
- Annual Budget
- -
- Monthly Burn Rate
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
- Fiscal Sponsor
- -
Theory of Change
Beijing-AISI believes that AI safety risks can be systematically identified, measured, and mitigated through rigorous evaluation benchmarks, technical safety research, and governance standards. By developing open evaluation frameworks (like ForesightSafety-Bench and PandaGuard) and publishing research findings at top academic venues, the institute aims to raise safety standards across the AI industry. Through policy engagement with China's national AI governance bodies and participation in international forums, it seeks to translate technical findings into governance norms and standards. The institute frames safety and governance as complementary: safety provides the technical guardrails to prevent harms, while governance steers AI development toward beneficial directions and prevents misuse. International cooperation and participation in multilateral AI safety initiatives are seen as key to ensuring global AI development remains safe and beneficial.
Grants Received
No grants recorded.
Projects
No linked projects.
People
No linked people.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.
Details
- Last Updated
- Apr 2, 2026, 10:00 PM UTC
- Created
- Mar 19, 2026, 10:31 PM UTC