Cohere chief calls for global AI safety body, with China at the table

Cohere chief executive Aidan Gomez is calling for a government-led global body to coordinate AI safety, including efforts to bring China into the conversation. His proposal pairs international coordination with independent testing of advanced models and a clear route for companies to report incidents.

Cohere Watch analysis

What happened

In an interview with the Financial Post, Gomez argued that AI safety should not be left to a handful of Silicon Valley companies. He said a public body should bring governments and technology companies together to coordinate standards and regulation, and that advanced models should face independent testing.

Gomez also called for companies to have a clear way to report when things go wrong, so incidents can be shared and learned from. He pointed to risks such as cyberattacks and bioweapons as tangible areas for attention, while stressing uncertainty about how the technology will develop. On international cooperation, he said working with China would be difficult but necessary. Read the Financial Post interview.

Why it matters

This is a concrete alternative to treating AI safety as either a private agreement between leading labs or a contest over who should slow down first. Gomez’s proposal puts public coordination, outside testing and incident reporting on the table, while arguing that global standards need more than a club of Western companies.

The challenge is in the machinery. A call for a global body does not settle who sets its rules, what evidence labs must share or how testing would work across borders. And including China may be easier to call for than to achieve. Still, the disagreement over who gets a seat is part of the governance question, not a footnote to it.

Our read

Gomez is making a case for building shared oversight around specific risks, rather than relying on a few companies to agree among themselves. That is a useful contribution to the debate, not proof that one global body can resolve it. The test is whether proposals for independent checks and incident reporting turn into practical standards that governments and labs will accept.

What to watch

• Whether governments or international organisations take up a formal coordination proposal.
• What independent testing would cover, and who would be allowed to conduct it.
• Whether companies develop a shared process for reporting AI-related incidents.
• Whether China joins any future discussions on standards or oversight.

Discussion spark: Should AI oversight be built through a global body that includes China, or through smaller groups of governments and labs that can agree on rules sooner?

not affiliated with or endorsed by Cohere