A broad federal system for overseeing frontier AI is unlikely to advance in the United States before the midterm elections. WIRED reports that a White House-linked regulatory proposal has stalled and that several congressional alternatives face both political opposition and a shrinking legislative calendar.
Treasury and White House science officials had drafted the outline of a self-regulatory body after Google DeepMind cofounder Demis Hassabis advocated such an approach. The effort was put on hold after technology executives objected and President Donald Trump hardened his opposition to AI regulation, according to people cited by the publication.
In Congress, the bipartisan Frontier Act would place auditors inside AI labs and allow the Commerce Department to intervene when a model presents catastrophic risk. It has support from some lawmakers and AI companies, but no clear route to a House vote. Separate proposals would give the government authority to shut down dangerous models, yet those bills also lack a consensus.
The practical result is a policy gap during a period of increasingly capable models and newly reported safety incidents. Narrow provisions could still be attached to must-pass legislation, but the source finds no single oversight proposal with enough White House and congressional support to become a comprehensive federal regime soon.