Xi tells White House AI must stay under human control

Chinese President Xi Jinping stood in the White House Grand Foyer on Thursday and said the United States and China have a responsibility to keep artificial intelligence under human control. President Donald Trump, speaking first, put “superintelligence, SI” on the list for their talks, hours after posting that he wants to leave AI policy where it is.

“Both China and the United States are leading nations in artificial intelligence,” Xi said through an English interpreter. “We have both the capability and responsibility to develop and manage AI for good, and ensure that the development of AI is always under human control and serves the well-being of the people.”

These were ceremony remarks, not an agreement. Neither leader announced a mechanism, a text or a test. The remarks came before formal talks.

Facts

Trump spoke first. He covered the two leaders’ history, trade, the Board of Trade set up during his May visit to Beijing, and farm market access. Then he turned to AI.

“During this visit, we will also discuss pressing issues concerning security, technology, and superintelligence, SI,” Trump said. “The decisions we make in these areas today can promote peace and prosperity for decades and decades to come.”

He did not define “SI” in the ceremony. Trump has been using “superintelligence” as his term for AI, The Hill and Fortune reported.

Xi’s AI lines came in a list of areas for cooperation, according to the official English text. They sat next to counternarcotics, law enforcement and more direct flights. Xi also said the two countries should keep competition within bounds.

Earlier Thursday, Trump posted on Truth Social that superintelligence would be a big topic but that he wants to leave it exactly where it is. He said that is also China’s position and called the Justice Department the guardrail, The Hill reported. He did not repeat those lines in the ceremony remarks.

Meaning

The two governments are using the same words for different plans. Beijing’s line is human control and “AI for good.” Washington’s line, from Trump’s post, is to leave AI policy where it is, with the Justice Department as the check.

Neither side offered anything in public on Thursday that can be verified. There are no tests to share, no incident thresholds and no named agency. Treasury Secretary Scott Bessent’s Sep 22 proposal for a U.S.-China AI incident-notification channel is the most concrete proposal on the public record, and neither leader mentioned it in the ceremony.

For context, neither government signed the 20-plus-nation call for frontier-AI control that led our edition this morning.

What’s next

• Any readout or joint statement from the talks that names AI, and whether it includes Bessent’s incident-notification idea.
• Whether the White House defines “SI” or links it to the AI Force and AI czar Trump announced Sep 21.
• China’s Foreign Ministry and Xinhua readouts, and whether they repeat the human-control line.