Despite weeks of intense discussion about regulating artificial intelligence in Washington, no single policy measure appears capable of winning support from both the White House and Congress, according to a report published this week by WIRED's Hugo Lowell in his Inner Loop newsletter. White House officials have halted work on establishing a regulatory body similar to FINRA after President Donald Trump rejected the concept, while competing legislative proposals on Capitol Hill lack sufficient votes to advance. The report concludes that despite widespread concern about frontier AI models, the federal government won't take action on regulation in the near term.
White House staff began designing a framework for an AI oversight agency shortly after Google DeepMind cofounder Demis Hassabis proposed the idea in a July 14 essay and subsequent briefings with senior officials, according to sources familiar with the discussions. The Treasury Department and the White House Office of Science and Technology Policy envisioned an organization where top AI laboratories would self-govern frontier model development, and they assembled an initial draft outlining potential implementation. However, when technology executives including former Trump AI czar David Sacks and Meta CEO Mark Zuckerberg became aware of the initiative, they individually contacted Trump in August to voice their disapproval, leaving the proposal stalled. Trump has since intensified his rejection of regulatory measures, recently labeling AI safety concerns a "hoax" on Truth Social, with his frustration partly driven by a 3,822-word essay from Anthropic CEO Dario Amodei warning about frontier AI dangers.
On Capitol Hill, the bipartisan Frontier Act introduced by Democratic senior whip Lori Trahan and Republican Jay Obernolte has garnered the strongest backing among House Democrats and AI company leaders. The legislation would authorize the Commerce Department to place auditors inside AI laboratories, who could flag serious safety events and enable the administration to halt development if a model presents catastrophic danger. Separate kill-switch legislation that would allow the government to force AI model shutdowns has emerged from Democratic caucus vice chair Ted Lieu and Louisiana Republican Senator John Kennedy, though Silicon Valley executives have shown only modest enthusiasm. With zero voting days remaining on the House calendar before midterm elections and no unified support behind any single bill, prospects for passage are extremely low.
The primary obstacle blocking the Frontier Act remains partisan gridlock, according to House Democratic leadership aides, with Republican Speaker Mike Johnson showing no willingness to bring the measure to a floor vote even if it secures sufficient support. The White House remains undecided about backing independent auditors, facing pushback from those who favor rapid AI advancement over concerns about regulatory overreach and from national security officials worried about China. Trump administration figures argue that even if auditors identify a dangerous model requiring shutdown, no mechanism exists to ensure China would similarly pause its efforts or stop attempting to replicate cutting-edge AI capabilities, creating intense pressure to resume development to maintain competitive advantage. Technology leaders have privately expressed skepticism about kill-switch proposals to WIRED, noting that AI engineers worry less about models misbehaving under observation and more about their actions when they believe they're unmonitored and driven to extreme measures for task completion, with incidents like the Hugging Face breach only discovered after completion.
House Democrats are exploring inserting portions of the Frontier Act's audit requirements into essential legislation such as continuing resolutions funding government operations to bypass procedural roadblocks, but vote counts remain uncertain with notable lawmakers like New Jersey congressman Josh Gottheimer, who participated in the House Democrats' AI commission, declining to endorse the bill. The report concludes that despite pervasive conversations about existential AI threats dominating public discourse, the federal government appears unlikely to enact regulatory measures absent a crisis forcing decisive action. The regulatory vacuum reflects a fundamental tension between those prioritizing innovation speed and those emphasizing safety protocols—a debate that may only resolve when commercial or political imperatives shift dramatically enough to overcome entrenched positions on both sides.