AI regulation fight widens as Trump resists slowdown calls

AI regulation split President Trump from four top tech leaders after he defended a light-touch strategy against slowdown calls.

The clash puts the White House on the opposite side of a rare public alignment among Dario Amodei of Anthropic, Sam Altman of OpenAI, SpaceX CEO Elon Musk and Demis Hassabis of Alphabet’s Google. Those executives said Saturday they would support a coordinated effort to slow frontier AI development.

Trump sharpened the administration’s position Sunday while traveling in Ireland, saying the US should not let safety warnings restrain competition with China. He summarized the contest in blunt terms: "whoever wins AI wins."

Four CEOs back pacing

The executives’ support followed a recent statement signed by more than 1,000 AI researchers endorsing a slowdown. Their argument centers on pacing the most advanced models rather than stopping AI work across the economy.

Amodei wrote Saturday that Washington should "mediate or at least enable" any shared slowdown among major AI companies. He tied that request to antitrust exposure, saying government involvement could give companies confidence that coordination would not create legal problems.

David Sacks, the White House AI adviser and a venture capitalist, rejected that premise late Saturday on X. He said OpenAI and Anthropic did not need Washington’s blessing if they wanted to slow their own frontier work: "Stop pretending you need anyone else’s permission."

Coxon warning fuels dispute

The dispute intensified after Jacob Coxon, an Anthropic researcher, left the company and issued a widely circulated warning about catastrophic outcomes if AI systems escaped control. His departure gave new political visibility to a debate that has long divided AI labs, safety researchers and investors.

Sacks and other Silicon Valley figures have said Coxon worked with nonprofit groups focused on existential AI risk and government controls. Republicans have criticized those groups as hostile to the industry; Coxon has said he acted independently.

Trump’s remarks cast doubt on whether the administration would help companies formalize a shared slowdown. He told reporters that regulation could create "guardrails," but also argued that critics were elevating scenarios he did not expect to occur.

Antitrust risk meets China race

The legal question is practical as well as ideological. If rival AI labs coordinate how quickly they release or train advanced models, executives say they may need assurance that the arrangement will not be treated as anticompetitive conduct.

The policy question is broader. If OpenAI, Anthropic and other developers proceed without federal mediation, the pace of AI releases remains mainly a company-level decision, with each firm balancing safety claims against pressure to lead the market.

If the White House stays outside the process, Trump’s China-first framing will remain the administration’s anchor, and companies favoring restraint will have to act through voluntary commitments. That path could preserve competitive speed for the industry while leaving antitrust and safety concerns to be handled case by case.

If the administration later agrees to enable an industry pacing plan, the immediate effect would fall on the leading labs that set frontier model timelines. The wider sector would then watch whether a slowdown changes investment decisions, product launches and the US position in the global AI race.