AI-generated editorial illustration; not a documentary photograph.
The White House has paused work on a proposed AI oversight body, while competing congressional bills lack the support and remaining floor time needed to advance, WIRED reported on September 16, 2026. The result leaves the US without a clear federal route for regulating risks from frontier AI models.
Oversight Agency Put on Hold
According to WIRED, White House aides began developing a framework after Google DeepMind cofounder Demis Hassabis proposed an oversight model in a July 14 essay and briefings with senior officials. The Treasury Department and Office of Science and Technology Policy drafted an initial concept under which leading AI laboratories would regulate frontier-model development through an administration-linked body.
The effort stalled after technology executives, including former White House AI adviser David Sacks and Meta CEO Mark Zuckerberg, contacted President Donald Trump in August to oppose it, the report said. WIRED's sources said no replacement oversight proposal was then under serious White House consideration.
Frontier Act Faces Political Barriers
The bipartisan Frontier Act would authorize Commerce Department auditors inside AI laboratories to report serious safety incidents. Its draft provisions would also allow the administration to halt development when a model presents catastrophic risk. OpenAI has publicly supported the measure, but WIRED reported that the White House remained undecided about independent auditors.
The bill also lacks a clear path through the House. Democratic leadership aides told WIRED that Speaker Mike Johnson was not interested in bringing it to the floor, while support among House Democrats was not unified. Some Democrats were considering attaching audit provisions to must-pass government funding legislation.
Shutdown Proposals Compete
Separate proposals from Representative Ted Lieu and Senator John Kennedy would give the government authority to force an AI model offline. Although the concept has bipartisan and bicameral support, WIRED reported only modest interest from Silicon Valley executives.
Timing Narrows the Options
No House voting days remained before the midterm elections at the time of the report, sharply limiting prospects for standalone legislation. Disagreement over audits, forced shutdowns and whether US restrictions could put domestic developers behind China further complicates efforts to assemble a governing coalition.
Source: Washington Won't Be Regulating AI Anytime Soon
SIGNAL TITANS
Leave a Reply