The Trump administration's approach on AI so far has been to prioritize innovation over regulation. But recently, there's been a shift in rhetoric.

AILSA CHANG, HOST:

The race between the U.S. and China on artificial intelligence is top of mind this week with President Trump in Beijing. So far, the Trump administration's approach to AI has been to prioritize innovation and root out any regulations around the technology. But there has been a shift in rhetoric from the president and others in his orbit. NPR White House correspondent Deepa Shivaram reports.

DEEPA SHIVARAM, BYLINE: President Trump hit the ground running on AI when he came back into the White House last year, and safety around the technology wasn't part of the equation. In his first trip abroad, Vice President JD Vance scolded European countries for their strict regulations around AI. He talked about it in an episode of the "All-In" podcast last year with then-White House AI czar David Sacks.

(SOUNDBITE OF PODCAST, "ALL-IN")

JD VANCE: America is done with the overregulating, with the constant worrying about the future. We're going to lean into the future. We're going to embrace the future and try to make sure America dominates in the future. And that's such a radical departure from both the safetyism of the Biden administration but also, frankly, the safetyism of a lot of our friends in Europe.

SHIVARAM: Within his first days, Trump revoked the Biden administration's executive order on creating safeguards around AI development. The administration even renamed the government's AI Safety Institute to the Center for AI Standards and Innovation. Since then, Trump has stomped on some state efforts to pass their own AI laws, saying a, quote, "patchwork" of regulation would be too complicated. But in recent weeks, the White House has been sending a different message.

(SOUNDBITE OF ARCHIVED RECORDING)

PRESIDENT DONALD TRUMP: There a lot of good things, but we have to be careful with it. Like everything else, we have to be careful with it.

SHIVARAM: That was Trump on Fox News a few weeks ago. In the same interview, he said, quote, "there should be regulations on AI." The shift in rhetoric comes after AI company Anthropic sounded the alarm about their latest model Mythos. The technology is powerful but too powerful to release to the public, Anthropic said, because of cybersecurity concerns. The White House soon jumped into action. White House Chief of Staff Susie Wiles met with Anthropic's CEO at the White House, and administration officials started warning banks of the risks that could come from technology like Mythos, using words that they haven't used before, like safety. Here's Treasury Secretary Scott Bessent in an interview on Fox News.

(SOUNDBITE OF ARCHIVED RECORDING)

SCOTT BESSENT: There is a very important calculus here between innovation and safety. And at the U.S. government, we're going to make sure that things stay safe.

SHIVARAM: The sudden concern over AI safety came as a surprise to Kyle Chan. He's an expert in China and tech policy at the Brookings Institution.

KYLE CHAN: Safety had become a bad word, and people started using words like AI security instead to try to get around that. And now, safety is front and center again. So it's a dramatic reversal from just a year ago.

SHIVARAM: Alondra Nelson, who led the Office of Science and Technology Policy in the Biden administration, says this rhetoric shift also comes as AI is starting to be a political problem for Trump. Polling shows Americans have become more skeptical about AI and want more regulations.

ALONDRA NELSON: The Trump administration is really looking ahead to the midterm elections and thinking about, you know, what might be done to sort of help the public see that these issues are being addressed.

SHIVARAM: But the ongoing AI race with China means there are still some in the administration who don't want regulations because they are concerned about slowing innovation and alienating tech leaders who Trump is cozy with. Still, the Trump administration is considering an executive order that addresses issues of safety and security. The White House has said any discussion of an executive order is speculative and any policy announcement would come from the president. But actual regulation of AI will have to come from Congress, though there's no indication of that happening anytime soon. Deepa Shivaram, NPR News, Washington.

(SOUNDBITE OF HI-TEK SONG, "ALL I NEED IS YOU (FEAT. CORMEGA & JONELL)")

Copyright © 2026 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

Accuracy and availability of NPR transcripts may vary. Transcript text may be revised to correct errors or match updates to audio. Audio on npr.org may be edited after its original broadcast or publication. The authoritative record of NPR’s programming is the audio record.