The Trump administration has moved away from the "light-touch" approach to artificial intelligence it started with and toward a far more interventionist stance — one that has produced restrictions on top AI models inside the United States. That's according to reporting by Leo Schwartz of The Information, surfaced by Techmeme.
The reversal is notable because the second Trump term opened with a very different signal. According to The Information, in January 2025 — just three days after Trump entered office for his second term — the president took early executive action on AI, a move widely read at the time as a promise to keep regulation minimal and let the industry move fast.
The new reporting traces how that early hands-off posture gave way to direct government intervention, to the point where some of the most capable AI systems now face restrictions in the U.S. market. The Information frames this as a genuine shift in philosophy rather than a one-off decision.
The full account of exactly which models are affected, and what specific policy levers drove the change, is laid out in The Information's story. The brief here reflects only what that reporting establishes: an administration that campaigned and began on the side of deregulation has, over the course of the term, embraced a more controlling role over advanced AI.
Why it matters: how Washington chooses to police the most powerful AI models shapes what tools American companies and consumers can access, how fast the technology spreads, and who ultimately sets the rules for an industry moving faster than the laws around it.