The Trump administration's internal divisions are freezing efforts to establish federal oversight of advanced artificial intelligence systems, leaving the AI industry and Congress in limbo weeks after Anthropic's latest powerful model rattled policymakers.
No new federal AI regulation has materialized since Mythos, Anthropic's most advanced model, captured Washington's attention. Early proposals for federal safety reviews have stalled as administration officials openly contradict one another on what, if anything, the government should actually do.
The timing is acute. Trump is heading to Beijing this week, and the issue of AI dominance is likely to surface in those talks. One tech industry source told Axios the administration may hold off on any major AI policy decisions until after the China summit concludes. The calculation appears to be that settling a position on advanced AI can wait until the diplomatic visit is done.
The friction among officials became visible last week when National Economic Council Director Kevin Hassett floated the idea of an "FDA for AI," comparing the approval process for new models to drug regulation. Within hours, White House chief of staff Susie Wiles and David Sacks, the co-chair of the President's Council of Advisors on Science and Technology, walked back his remarks. Hassett himself later told CNBC that creating "a giant new bureaucracy to approve AIs" was not being considered.
Sacks has instead emphasized a different concern: the threat from foreign AI development. He told Fox Business that the real danger lies not in what American labs are building but in advanced models being trained by China and other actors, which could gain dangerous cyber capabilities within months. His focus centers on hardening U.S. systems against those risks rather than regulating American companies.
The divergence reflects a deeper disagreement about the federal role. Another tech industry source close to government discussions told Axios that "there are some differing views in the administration about how to approach the issue of these highly capable AI systems." Industry players are requesting clarity on how the government will handle future breakthroughs, signaling frustration that policy is being made reactively rather than proactively.
Behind the scenes, another battle is taking shape over which part of government should oversee advanced AI testing. Some officials favor a civilian approach through the Department of Commerce, while others push for national security agencies to take the lead. That tug-of-war came into sharp focus when the Commerce Department's Center for AI Standards and Innovation publicized new testing deals with frontier AI companies. Days later, the website was taken down. Staff were instructed to remove the page with no explanation provided. The White House and Commerce have declined to say why the page was pulled or whether the program continues.
Rep. Josh Gottheimer, a New Jersey Democrat who co-chairs the House Democratic Commission on Artificial Intelligence and the Innovation Economy, warned against delay. "We don't have time to waste, to point fingers or to let bureaucracy get in the way," he told Axios, calling for executive action on testing and evaluation of new models while Congress pursues legislation.
Industry sources continue to expect new guidance on AI cybersecurity, model testing, and contractor rules, but the window for action may be narrowing. Administration infighting, combined with the demands of the China trip and other competing priorities, threatens to push AI policy further into the background.
Author James Rodriguez: "The administration's public squabbling over whether to regulate AI at all signals deeper confusion about national strategy on the technology that matters most."
Comments