The Trump administration has withdrawn an executive order requiring AI safety testing after major technology company chief executives boycotted a White House event designed to showcase the policy, according to Ars Technica. The cancellation represents a significant retreat on federal AI oversight following coordinated industry pushback.
The executive order, which would have mandated safety testing protocols for advanced AI systems before deployment, was shelved after fewer than half of invited tech CEOs attended the planned White House announcement ceremony. The administration had positioned the event as a major policy milestone, but the sparse attendance forced officials to reconsider the directive’s viability.
According to sources familiar with the matter, the boycott reflected broader industry opposition to prescriptive federal testing requirements. Technology executives reportedly objected to compliance costs and timeline constraints that would have applied to AI models exceeding specified computational thresholds. The administration had not publicly disclosed the technical parameters that would trigger mandatory testing before the order’s cancellation.
The policy reversal marks the second major AI regulatory initiative abandoned by the Trump administration within six months. The administration previously withdrew proposed export controls on AI chip technology after semiconductor industry lobbying. These retreats stand in contrast to the European Union’s AI Act, which entered enforcement in August 2024 with mandatory risk assessments for high-impact systems.
Industry sources indicated that chief executives from at least three of the five largest US technology companies by market capitalisation declined White House invitations. The coordinated absence effectively signalled that major AI developers would not voluntarily commit to federal testing protocols without congressional legislation establishing clear legal frameworks.
The business implications are substantial. Without federal testing mandates, AI companies retain flexibility to establish internal safety protocols and deployment timelines. This regulatory vacuum advantages established players with resources to conduct voluntary safety research whilst potentially creating barriers for smaller competitors seeking to differentiate through safety certifications.
However, the policy void leaves liability questions unresolved. Without standardised testing protocols, companies deploying AI systems face uncertain legal exposure if models cause harm. Insurance markets for AI liability remain underdeveloped, and the absence of federal safe harbour provisions may ultimately prove more costly than compliance with testing requirements would have been.
The cancellation also shifts competitive dynamics internationally. Chinese AI developers operate under government-mandated safety reviews before public deployment, whilst EU companies must comply with the AI Act’s transparency and testing requirements. US companies now occupy a regulatory middle ground—facing fewer domestic constraints but potentially encountering market access barriers in jurisdictions requiring safety certifications.
Congressional response will determine whether federal AI oversight emerges through legislation rather than executive action. Senator Maria Cantwell’s bipartisan AI safety bill, currently in committee, would establish voluntary testing frameworks with liability protections for participating companies. The executive order’s failure may accelerate legislative efforts by demonstrating that industry will not accept unilateral executive mandates.
Market analysts noted that AI infrastructure providers and cloud computing platforms may benefit from the regulatory uncertainty, as companies seek third-party safety testing services to manage legal risks voluntarily. Specialist AI safety firms have reported increased enterprise interest since the executive order’s cancellation.
The immediate outlook centres on whether the administration attempts modified oversight approaches or abandons prescriptive AI regulation entirely. Industry observers will monitor upcoming congressional hearings and whether the Commerce Department pursues safety standards through the National Institute of Standards and Technology’s AI Safety Institute, which operates independently of executive orders.
The episode underscores the technology industry’s capacity to block federal AI oversight through coordinated opposition, raising questions about whether effective regulation can emerge without legislative mandates that companies cannot simply boycott into irrelevance.
