OddsFrontTrump rejects calls for an AI slowdown and makes competition with China his priority
The president’s position sharpens the divide between the White House and technology executives who now argue that safety controls need time to catch up with frontier models.
The White House chooses speed
US President Donald Trump rejected calls to slow the development of advanced artificial-intelligence systems during his visit to Ireland on September 13. Responding to warnings from researchers and technology executives, he argued that many feared outcomes would not occur and placed greater emphasis on preserving the United States’ lead over China. The remarks establish a clear presidential position at a moment when parts of the AI industry are asking for more time to strengthen safeguards.
Trump did not rule out guardrails altogether, but he presented an American slowdown as a strategic risk if Chinese laboratories continued advancing. That framing is consistent with his administration’s established policy: promote rapid domestic innovation, infrastructure and exports while addressing specific harms through narrower rules. The White House’s legislative framework recommends protections for children, law-enforcement action against AI-enabled fraud and stronger national-security expertise, while avoiding broad licensing requirements.
Industry pressure is moving in the opposite direction
The president’s intervention followed a coordinated change in tone from leading US laboratories. Axios reported that four major AI companies endorsed a slower pace for frontier-model development after warnings that safety work and public oversight were falling behind capabilities. Their proposals differ in detail, but they share a demand for independent evaluation, stronger coordination and more time before systems reach potentially critical levels of autonomy or cyber capability.
The disagreement is not simply between regulation and deregulation. Both the administration and leading companies acknowledge concrete risks involving children, fraud, cyber operations and national security. The dispute concerns how much weight to place on low-probability catastrophic scenarios, whether voluntary commitments can be trusted, and whether binding constraints would hand an advantage to competitors operating outside a common framework. China has become the central argument against unilateral restraint.
Attention now turns to Congress and any meeting between Trump, lawmakers and technology executives. A durable policy would have to reconcile three objectives: keeping US laboratories competitive, creating credible mechanisms for testing powerful models, and preventing a regulatory system that protects only the largest incumbents. Trump’s rejection of a general slowdown makes an immediate nationwide pause unlikely, but it leaves room for narrower rules already contemplated in the White House framework.