White House to AI labs that fear their models: 'just stop it'

The White House has answered the AI chiefs who are asking for a slower pace, and the answer is: do it yourselves. On Fox News Sunday on September 20, Michael Kratsios, the president's science and technology adviser, said a company that believes it is building something unsafe "can just stop it."

He spoke a day after President Trump announced on Truth Social that he would create an "AI Force," which he likened to the Space Force of his first term, and would soon appoint an "AI Czar."

What Kratsios said, and what he left out

The target was plain. Anthropic's Dario Amodei recently published a blog post calling for a slower pace of AI development, and OpenAI's Sam Altman said his company would welcome one, ABC News reports. Kratsios's reply was that the executives "just need to slow down if they feel that they need to do so." Regulation, he said, should be "use case and sector specific," handled by the agencies that already oversee drones, self-driving cars and medical diagnostics.

He offered one example of the government acting. This summer, he said, "a leading model developer" released a model, the government received credible evidence of cyber issues with it, a letter was sent, and the model was pulled down until the problems were resolved. He did not name the developer or the model. Nor, in the remarks as reported, did he say which agency sent the letter or under what authority.

There is an obvious objection, and it is ours rather than anything said on air. A lab that pauses alone hands the lead to a rival and changes little else, which is presumably why the executives ask for rules that bind everyone. The quoted remarks do not deal with it.

An 'AI Force' without a job description

Trump's post, as quoted by CBS News, reads: "We will not in any way hinder or stifle the Growth of this incredible Industry. Rather, we will cherish it, help it, and watch over it, as it grows!" Then the other half: "However, we will also be looking for BAD, and we can do that, very easily, with our already existing Criminal and Civil Justice System." He called AI "the next Industrial Revolution, or Internet," only bigger, "possibly as much as 25% of our Country's GDP," and dismissed fears of a superintelligence that threatens humanity as a hoax.

What would an AI Force do? Nobody has said. ABC News reported that its duties are unclear, that a new branch of the armed forces would need an act of Congress, and that the White House did not immediately comment. The czar is a revival. David Sacks was the AI and crypto czar until March and now co-chairs the president's council of science and technology advisers. His line matches Kratsios's: "I think the starting point for talking about AI regulation is to realize the regulations that we already have."

The Sunday shows did not split on party lines

Nvidia's Jensen Huang went further on CBS Sunday Morning. "Read between the lines," he said of the lab chiefs. "They're actually not asking for more laws. They're asking to be relieved of the laws we do have." Product liability already covers a product that does harm, he argued: "We have plenty of laws, so let's apply those first." Bear in mind who is speaking. Nvidia sells the chips these labs run on.

In Congress the argument cut across parties. Senator Dave McCormick, a Pennsylvania Republican, said on Fox News Sunday that executives cannot raise "hundreds of billions of dollars" for their businesses and "at the same time, say it poses a risk of destroying humanity." Senator Peter Welch, a Vermont Democrat, agreed that industry needs standards but added: "I think there has to be a public role." On NBC's Meet the Press, Senator Cory Booker of New Jersey said: "Call Congress back in."

Who gains from 'use the laws we have'

If the administration holds this line, the companies that gain are those that want to keep shipping: chip suppliers, cloud providers and any lab that does not share the alarm. The labs asking for a common rule lose the cover it would have given them. They can still slow down, as Kratsios says, but only by accepting that competitors may not.

The public is left with existing criminal and civil law, which acts after harm has been done. Whether that is adequate for AI systems is exactly what the two sides disagree about, and the weekend of September 19 and 20 did not move either of them.

What to watch

Still missing are the czar's name and any document that says what the AI Force is or where it sits in government. Fox News reported that Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng were due to discuss AI "guard rails" on Sunday, ahead of a meeting between Trump and Xi Jinping. Whether AI appears in whatever the two leaders agree will say more about how the administration weighs the risk than a weekend of television did.