OpenAI publicly backed the creation of an international AI oversight body on Wednesday — led by the US but with China as a member. The model: the International Atomic Energy Agency (IAEA). The timing: hours before the Trump-Xi Jinping meeting.
The proposal
Chris Lehane, OpenAI’s VP of Global Affairs, argues that the US should use its AI lead to create a global governance mechanism. The idea isn’t new — an AI-IAEA has been debated for years — but OpenAI officially getting behind it carries weight.
The core of the proposal: an international body should ensure safer, more resilient AI systems. Similar to how the IAEA conducts inspections and sets standards for nuclear energy, an AI agency would establish evaluations, safety standards, and cooperation mechanisms.
Meanwhile: US-China AI talks
On the same day, US Treasury Secretary Bessent met with Chinese representatives to discuss AI rules. Bessent emphasized that the US leads in AI — and that a safety protocol is planned.
Anthropic also weighed in: with a policy paper warning about China’s potential to catch up in AI and calling for tighter chip controls. Two of the world’s biggest AI companies pushed AI geopolitics center stage on the same day — with different strategies.
Context
OpenAI’s move is also an image offensive. While the company is simultaneously in court with Musk, dealing with Apple troubles, and having to explain its path from nonprofit mission to for-profit giant, it’s positioning itself as the voice of reason in global AI policy.
Whether this actually leads to an international agency is another question entirely. But the fact that the biggest AI labs are now actively making governance proposals shows something: the industry knows regulation is coming — and would rather help shape it than just react.
Sources: Japan Times, Seeking Alpha, CNBC