US and China discuss AI safety protocol as rivalry continues

In short: The United States and China say they plan to talk about AI safety, but the timing and details are still unclear.

What's going on

U.S. Treasury Secretary Scott Bessent said the U.S. and China are starting a new set of talks about artificial intelligence, with a focus on safety and security. He spoke while in Beijing during high-level meetings involving President Donald Trump and China’s leader Xi Jinping.

Bessent said the two countries aim to create an “AI safety protocol.” A protocol is a shared set of rules or steps, like a checklist both sides agree to follow. He also talked about “best practices,” meaning practical guidelines for how powerful AI systems should be handled.

He did not say exactly when the talks would happen or how often officials would meet. Public reports so far describe the talks as connected to the current summit, but they do not lay out a fixed schedule.

At the same time, Bessent made clear the U.S. wants to keep its advantage in AI. He pointed to U.S. limits on selling certain advanced computer chips to China, since those chips are important for training and running modern AI systems (like the engine parts needed to build faster cars).

What to watch

Watch for whether the two governments announce a formal process, such as regular meetings, named working groups, or specific topics like stopping criminals and other non-government groups from getting access to very powerful AI models. Also watch whether safety talks reduce tensions, or whether the competition over chips, trade, and military uses keeps escalating even as officials say they want shared rules.

Source: NYTimes