Huawei's Eric Xu: Chinese AI not yet powerful enough to observe frontier risks
sunychoudhary · reddit · 2026-09-17
According to Reuters, Huawei's Eric Xu argued that Chinese AI labs may not yet operate at a capability level where they can observe the same frontier risks reported by U.S. labs.
- His implicit logic: some safety problems only become visible once systems are sufficiently capable, not equally obvious at every stage
- This poses a hard question for AI safety: if certain failure modes only appear near the frontier, how can labs know which safeguards to deploy in advance?
- It also complicates international coordination, as different countries may assess AI risk from very different capability levels
Related event: Huawei's Xu Zhijun Says China's AI Must Speed Up to See Its Dangers(4 posts)→
More from AGI Musings
- Forward Deployed Verifier will soon be as valuable as FDE, argues poster — vaibhavbetter · 2026-11-02
- 'AI is a massive balance patch and elo reset in the game of life—and everyone's still playing the old meta' — zack_overflow · 2026-11-02
- Why AI isn't 'alien intelligence': a tree is stranger than any model — LucaAmb · 2026-11-02
- Mathematician Tony Feng: current-paradigm LLMs lack the diversity of thought research needs — littmath · 2026-11-02
- METR's independent probe of the OpenAI agents' multi-day Hugging Face hack — gleech · 2026-11-02
- MIT Tech Review: The agentic web is here — AI agents are remaking the internet — MIT Tech Review AI · 2026-11-02