Trump rules out slowing US AI development after his summit with Xi

A federal brake on AI can come from two places, the president or Congress, and on Saturday the president gave his answer ^[1]. He put the case for slowing down in the same category as climate change and the inquiry into whether his 2016 campaign colluded with Russian operatives ^[12]. "When they wanna put on brakes, it's just like everybody's going to be dead. You're all supposed to be dead now, because of global warming. It's like Russia, Russia, Russia hoax. All of these different hoaxes," Trump said ^[13]. He gave a competitive reason. "What they wanna do is stop our progress, because we're leading China by a lot," he said. "And we're going to keep it that way." ^[2]

The summit shows what the administration chose to spend time on. US lawmakers and experts had hoped the three-day meeting would produce a breakthrough AI agreement. It did not ^[6][11]. "We didn't spend that much time talking about it," Trump said ^[8]. He indicated that what AI discussion there was focused mostly on his preferred name for the technology, "super intelligence" ^[16]. China's foreign ministry said on Saturday that it "respects" the switch, which Trump announced on Tuesday for official communications ^[15]. He also ruled out working with Beijing on the technology. "I would rather not integrate, because we're leading by a lot," he said. "When you're leading, you don't open it up to each other." ^[7] Xi has said AI must develop under human control ^[4].

That covers half of the argument that federal rule risk for AI spenders has fallen. The executive-branch half is now on record in the president's own words ^[1]. Al Jazeera's report does not mention state legislatures or the midterm elections, and it cites no pending bill. So the claim that the risk moves to statehouses or to the next Congress is a forecast, and this evidence cannot confirm it. The lawmakers who want a slower pace exist ^[3]. The report gives no vehicle or date for anything they might pass.

A few events could change the picture. Any AI incident reported through the new bilateral channel would give those lawmakers a specific event to legislate around. Beijing had not commented on the channel when the White House announced it on Friday ^[9]. Xi is due to meet Trump two more times this year ^[14], three meetings in all counting the one just ended ^[17]. Each one gives Beijing another chance to press its line on human control, and its spokesperson has left the door open: "Artificial intelligence technologies are continuously developing, all parties can strengthen exchanges, discuss in-depth and seek consensus in light of the latest situation" ^[5]. Outside pressure is also on the record. Guterres warned in his final General Assembly address that development "without accountability" could put the world at risk ^[10].

In my view, none of these produces federal limits on AI development before the year ends, given how firmly Trump rejected brakes on Saturday. The view is wrong if the White House restricts development by order, or if either of the two remaining Xi meetings ends with a signed AI agreement that covers more than incident reporting ^[14].