US and China Will Begin AI Safety Talks, Bessent Says

The United States and China will begin formal talks on AI safety guardrails, US Treasury Secretary Scott Bessent said during President Donald Trump’s summit with Chinese President Xi Jinping in Beijing.

The discussions would be the first bilateral AI safety framework between the two countries and aim to establish a protocol that keeps the most powerful AI models out of the hands of nonstate actors, Bessent told CNBC on Thursday. Neither Washington nor Beijing offered a timeline, and it is not clear whether any protocol would be binding or voluntary.

“The two AI superpowers are going to start talking. We’re going to set up a protocol in terms of how do we go forward with best practices for AI to make sure nonstate actors don’t get a hold of these models,” Bessent told CNBC’s Joe Kernen.

Bessent said the talks were possible only because the US is ahead of China in the race to build advanced AI systems. “The reason we are able to have wholesome discussions with the Chinese on AI is because we are in the lead,” he said. “I do not think we would be having the same discussions if they were this far ahead of us.”

He told The New York Times that Washington planned to export its own regulatory model rather than negotiate a compromise with Beijing: “So we’re going to put in U.S. best practices, U.S. values on this, and then roll those out to the world.”

Bessent also acknowledged the trade-off between safety rules and the speed of commercial AI development. “What we don’t want to do is stifle innovation. So our responsibility is to come up with the highest performance calculus where we can get the most innovation and the highest level of safety,” he told Reuters.

The announcement came as Trump met Xi in Beijing. Nvidia chief executive Jensen Huang joined the American delegation, Techmeme reported. Nvidia’s most advanced chips are still subject to US export controls that limit sales to China, and any AI safety framework the two countries agree on would almost certainly overlap with existing technology transfer restrictions.

For Australia, a close US security ally and China’s largest trading partner, a bilateral AI safety deal between Washington and Beijing has real implications. Australian tech firms that operate in both markets would face a regulatory landscape shaped by whatever protocol comes out of the talks. The Albanese government’s own AI safety agenda — including mandatory guardrails for high-risk AI now under consultation by the Department of Industry, Science and Resources — could be drawn toward the US-led standards Bessent described, rather than the EU’s more prescriptive AI Act model that Canberra has so far watched from the sidelines.

Neither the White House nor China’s foreign ministry has commented on the proposed scope or timeline. Bessent’s remarks are still the clearest signal yet that the Trump administration sees AI governance as a domain where it can dictate terms to China, rather than one that needs mutual concession.