Despite growing concerns over the risks posed by highly capable AI models, meaningful federal regulation appears unlikely in the near term, with the White House firmly opposed to broad oversight and Congress divided over competing proposals.
For weeks, policymakers in Washington have debated whether the US needs a new regulatory framework for frontier AI systems. But efforts to establish an oversight mechanism have stalled, while lawmakers have yet to unite behind legislation capable of passing both chambers.
White House oversight plan put on hold
White House officials initially explored the idea of creating a FINRA-style body that would oversee the development of advanced AI models. The proposal emerged after Google DeepMind co-founder Demis Hassabis raised the idea in July.
According to people familiar with the discussions, the Treasury Department and the White House Office of Science and Technology Policy considered a framework under which leading AI companies would play a role in regulating the development of frontier models.
The effort lost momentum after several technology executives, including former Trump AI adviser David Sacks and Meta CEO Mark Zuckerberg, reportedly contacted President Donald Trump to voice their opposition.
No new major AI oversight initiative is currently being seriously considered by the White House, according to sources cited by WIRED. Trump has repeatedly rejected calls for stricter AI regulation and recently described concerns about AI safety as a “hoax.”
Congress faces its own hurdles
With the White House reluctant to act, lawmakers have introduced several competing proposals. But the chances of Congress passing comprehensive AI regulation remain uncertain, particularly with the House schedule leaving little time for legislation before the midterm elections.
One of the proposals receiving attention is the bipartisan Frontier Act, introduced by Democratic Rep. Lori Trahan and Republican Rep. Jay Obernolte.
The bill would allow the Commerce Department to place independent auditors inside AI companies. Those auditors could report serious safety incidents and potentially recommend halting the development of models considered to pose catastrophic risks.
OpenAI has publicly supported the proposal, but the White House has yet to endorse it. Some technology industry figures have raised concerns about excessive regulation, while national security officials have questioned whether restrictions could put the US at a disadvantage against China.
Kill-switch proposals gain limited traction
Another approach under consideration would give the government the power to force an AI model to shut down if it poses a serious threat.
Two competing proposals have emerged, including one from Democratic Rep. Ted Lieu and another from Republican Sen. John Kennedy.
However, technology executives have shown limited enthusiasm for the idea. A central concern among AI researchers is that dangerous behavior could occur before authorities are aware of it, making a shutdown mechanism ineffective if a system has already caused damage.
For now, Washington appears unlikely to impose broad federal oversight of frontier AI. Unless a major incident changes the political calculus, efforts to regulate the technology are likely to remain stalled.
Be First to Comment