Well, that didn’t take long. Days after Anthropic’s Dario Amodei warned that swarms of rogue AI agents could seize control of the internet within six months and begged the industry to slow down, President Trump swatted the whole thing aside. His message, in essence: stop scaring everyone, the real priority is beating China, and the catastrophe scenarios “won’t happen.” The most powerful man in the country just told the most nervous men in tech to get a grip.
It is a genuinely important collision, because it pits the two forces that will actually decide how AI gets governed against each other: the safety-worried labs and a White House that treats AI primarily as a weapon in a great-power race.
Why Trump’s framing matters more than the essays
Amodei, Altman, Musk and Hassabis can call for a slowdown all they like, but a voluntary pause means little if Washington’s stance is “full speed, beat Beijing.” Trump’s position reframes AI safety as a national-security liability: any brake the US applies to itself, the argument goes, is a head start handed to China. That is a powerful, simple message, and it lands with a political base that is not losing sleep over hypothetical superintelligence.
It also exposes the awkward corner the labs have painted themselves into. Having spent months insisting their technology might end civilisation, they now need a government to impose limits none of them will accept unilaterally. And the government they need has just told them, in front of everyone, that it thinks they are crying wolf. You cannot lobby for rules while your own product is your best evidence that the rules would cost you the race.
The takeaway
Here is the uncomfortable truth for anyone hoping the “slowdown” talk leads somewhere. Real limits on AI would need governments to enforce them, and the US government just made clear it is more worried about losing to China than about the doom scenarios its own AI CEOs are describing. That gap between what the labs say they fear and what the state is willing to do about it is the actual story of AI policy right now, and it is not closing. Watch the White House, not the essays. (Not investment advice.)
Related: the industry-wide slowdown call and the extinction-risk debate.