White House Weighs Pre-Release Vetting for AI Models
WASHINGTON — The White House is considering requiring artificial intelligence models to undergo pre-release vetting before public deployment, The New York Times reported Sunday.
The proposal, if enacted, would impose pre-release review requirements on AI laboratories before they could deploy models publicly, according to the Times report. The policy would affect major American AI companies including OpenAI, Anthropic, Google and Meta.
A mandatory pre-release vetting regime would mark a departure from the current U.S. regulatory posture, which has largely relied on voluntary commitments from AI companies and sector-specific guidance rather than binding pre-market review requirements.
The consideration comes as governments worldwide grapple with how to oversee a technology that is advancing faster than regulatory frameworks can adapt. The European Union’s AI Act, which begins enforcement of its broadest provisions in August 2026, takes a risk-based approach that includes conformity assessments for high-risk AI systems — but the U.S. has until now resisted comparable pre-market mandates at the federal level.
Industry groups have generally pushed back against pre-release review requirements, arguing they could slow innovation and put American companies at a competitive disadvantage relative to foreign rivals, particularly Chinese AI labs that operate under different regulatory constraints.
Proponents of pre-release vetting contend that the rapid deployment of increasingly powerful AI systems without independent review poses risks to national security, public safety and economic stability that voluntary commitments alone cannot address.
The details of the proposal — including which agency would conduct reviews, what thresholds would trigger vetting and how long the process would take — remain unclear from the initial report. Those specifics would likely determine whether such a policy functions as a meaningful safety check or a significant bottleneck on AI development.
Any executive action on pre-release AI vetting would also face questions about its legal durability, as Congress has yet to pass comprehensive federal AI legislation that would provide a statutory foundation for such requirements.
