China Says Trump and Xi Back More AI Dialogue, but No Specific Deal Is Confirmed

The United States and China leave this week’s leaders’ meeting with a shared call for AI dialogue, according to Beijing, but no confirmed agreement on what either government will do next. China’s account of President Donald Trump’s talks with President Xi Jinping describes cooperation on AI risks. It does not turn that language into specific commitments.

What Beijing put on the record

Trump and Xi met at the White House on Thursday. In China’s official readout, Xi argued that the countries could cooperate on AI even while competing in the technology. He called for exchanges about its risks and benefits and joint efforts to prevent misuse. He also said AI should remain under human control.

The same readout attributes a shorter position to Trump: the two countries should maintain dialogue and strengthen cooperation on AI. That is China’s account of his remarks, not a joint statement. The U.S. government had not issued its own summit readout when CNBC published its account of the talks.

Before the meeting, Trump wrote that “Super Intelligence” would be a major topic but that he wanted to “leave it exactly where it is.” The summit account is more explicit about talking together than about changing course. Neither leader’s reported remarks specify a new restriction on AI development or a procedure for handling a dangerous incident.

The two sides can continue their dialogue on AI, exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI.

Xi Jinping, according to China’s Foreign Ministry readout

An incident channel was already under discussion

The leaders’ remarks followed talks over the preceding weekend between U.S. Treasury Secretary Scott Bessent and his Chinese counterpart, He Lifeng. They discussed a U.S.-China AI dialogue and an incident channel covering situations up to the national-security level, CNBC reported. That earlier discussion gives the summit’s broad language a possible practical focus, though the summit did not confirm an operating channel.

The distinction matters for a channel meant to handle serious incidents. A willingness to exchange views does not say what would trigger a notification, who would send it or how the other side would respond. Those are questions a working arrangement would have to settle; they are not provisions announced in the Chinese summit readout.

Aalok Mehta, director of the Wadhwani AI Center at the Center for Strategic and International Studies, told CNBC an emergency channel could become useful if paired with an incident-reporting system. His assessment points to a test beyond getting the leaders to endorse dialogue: whether officials build a process they can use when a serious concern arises.

Cooperation beside a chip dispute

The AI discussion took place alongside a dispute over access to the hardware used to build advanced systems. U.S. export controls restrict China’s ability to import advanced Nvidia chips, a potential sticking point CNBC identified ahead of the talks. Xi’s appeal to draw on each other’s strengths, rather than guard against each other, sits beside that unresolved tension. There is no confirmed summit outcome on the chip issue.

The state dinner made the technology industry highly visible. Nvidia CEO Jensen Huang and OpenAI CEO Sam Altman were among the executives who attended. Their presence shows the prominence of AI around the visit, but does not give them a stated role in the presidents’ closed-door discussion or supply terms for an agreement.

For now, the public result on AI is a diplomatic position, not a set of operating rules. The Guardian found no sign of a breakthrough from the closed-door meeting on AI and other divisive issues. Any later agreement on incident alerts would need to answer the practical questions left open by Thursday’s language.