President Trump struggles to make a plan for artificial intelligence regulation after initially taking a hands off position on the emerging technology.

MARY LOUISE KELLY, HOST:

There has been some confusion in the White House over how the Trump administration wants to handle artificial intelligence. This week, President Trump was supposed to sign an executive order addressing the rapidly developing technology, but then he abruptly decided not to move forward with it. The decision comes as cybersecurity concerns over some AI models are rising. NPR White House correspondent Deepa Shivaram is following this. Hey, there. Hey there.

DEEPA SHIVARAM, BYLINE: Hi, Mary Louise.

KELLY: So what was this executive order supposed to accomplish? And why did President Trump pull back at the last minute?

SHIVARAM: Yeah, well, the president was set to sign this executive order yesterday afternoon, and it was meant to address some of the growing concerns about security and safety over AI. For weeks, White House officials have been talking about what that might look like - right? - because remember, Congress hasn't passed any regulation around AI. So this technology is rapidly developing, like you said, without a lot of guardrails.

But instead of signing the order yesterday, Trump said that he had concerns with some aspects of the order and how it might impact AI innovation. And his concern is about making sure that the U.S. stays ahead of China and other countries when it comes to the development of AI.

(SOUNDBITE OF ARCHIVED RECORDING)

PRESIDENT DONALD TRUMP: We're leading China. We're leading everybody. And I don't want to do anything that's going to get in the way of that lead.

SHIVARAM: And I reached out to the White House, and they did not respond to requests for comment. So it's not really clear at this point when the White House plans to reschedule the signing of this order or if the order will change in nature at all.

KELLY: And why has this administration - which, I should say, has been against AI safety measures all along - why are they even contemplating making moves on security now?

SHIVARAM: Yeah, I mean, that's a great point. Up until a few weeks ago, the Trump administration was pretty opposed to talking about AI safety. In Trump's first days back in office, he scaled back the Biden administration's executive order on AI that put in some regulations around the technology. And the Trump White House even changed the name of the AI Safety Institute to take out the word safety, and they've been largely opposed to any states passing their own AI regulations. So this was - you're right - not something that they wanted to talk about. And the administration's perspective was that any regulation or rules around safety would hamper innovation and the focus had to be on beating China.

KELLY: And what happened? What changed?

SHIVARAM: Right. So then a few weeks ago, Anthropic, which is one of the big AI companies, said that one of their new AI models, Mythos, was basically too powerful to release because of cybersecurity concerns, and it set off some alarm bells in the administration because what if technology like that was obtained by bad actors? Or what if it's used against the U.S., like attacks on banking systems?

So what followed was, like, this flurry of movement from the Trump administration. Officials started to talk about AI and use words like safety for the first time. Trump even said himself in a Fox News interview that there should be regulations around AI. And all of that back and forth was part of this debate and figuring out what the White House would do next. And ultimately, it was supposed to be this executive order, which would have laid out some ways to try and work with tech companies around these high-risk models.

But keep in mind, Mary Louise, it's not like the Trump White House is doing a full 180 here. Trump is still really cozy with tech leaders and tech companies and wants to generally keep a very light touch when it comes to setting up any kind of guardrails.

KELLY: Well, and meanwhile - and just briefly - to note, polling shows more and more Americans favor some kind of regulation of AI, so more and more voters. Is the White House taking that into account?

SHIVARAM: I mean, there was a recent survey out of the University of Pennsylvania that showed that Americans' view on AI is pretty negative overall. Sixty-five percent said the government has done too little to regulate AI. And polling like that shows how concerned Americans are over this technology, right? But I will say, politically, it's a pretty wide-open space for either party to jump in on, and so far, they haven't.

KELLY: NPR's Deepa Shivaram, thank you.

SHIVARAM: Thank you.

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.