Deadlock in Washington as Trump Dismisses AI Safety Fears as a Hoax

When President Donald Trump addressed mounting calls for a slowdown in artificial intelligence development on Monday, he was unequivocal. "AI taking over the World, destroying Humanity, and all other things bad, is a HOAX," he wrote on social media. He had earlier argued that the only control or guardrails the technology requires is a strong and smart president.

That view is not shared by senior Democrats in Congress. Nor is it endorsed by leading figures in the AI sector, who have in recent days issued stark safety warnings and made the case for government regulation. In Washington, however, new safety legislation on AI is unlikely anytime soon. Any bill would have to pass both the House of Representatives and the Senate to become law. Trump's staunch opposition to fresh guardrails means the Republican Party, which holds a majority in both chambers, is unlikely to break from him and back restrictions or government controls. Even if it did, whatever bill emerged would most likely require the president's signature to take effect. His dismissal of the need for regulation on Monday effectively hit pause on any hope for executive action.

Beyond this, a more immediate obstacle to any movement in Congress is the clock. The House is set to enter recess at the end of the week until after the midterm elections in early November. Some leading Democrats are calling for that recess to be delayed until AI safeguards can be passed. "Congress should not leave town until something is done decisively to protect the safety and the wellbeing of the American people," Democratic House Minority Leader Hakeem Jeffries told reporters on Monday. A group of House Democrats calling for more time to debate the issue have also written to Republican Speaker Mike Johnson, who would ultimately decide on delaying the recess. "Reasonable minds may disagree about precisely how Congress should regulate this rapidly evolving technology," the group said. "We cannot disagree about the imperative for Congress to act." Over the weekend, Republican congresswoman Anna Paulina Luna suggested on X that Congress should convene a special session on the issue.

It is highly unlikely, however, that the recess will be delayed or a special session convened in the coming days. Speaker Johnson, who on Monday faced a barrage of questions about AI from reporters in the halls of Congress, gave no sign he intended to move quickly this week. Asked whether the House would take action on AI concerns before breaking for the midterms, he replied: "This is a very complex issue." He said Trump would hold a meeting at the White House soon with AI executives to discuss the responsibility of companies to maintain safety and what role, if any, the government has to play. "We want to emphasise the necessity of [AI companies] providing safe products and self-regulating," Johnson said, before adding: "There is also a role potentially for Congress to play." What that role could look like is unclear, and it remains a contested question among those in Congress.

While Democrats are pushing for more time to address AI safety, they are themselves divided on how best to regulate the technology. Even if the party were to win control of both chambers in November's elections, there are no guarantees of swift action, as it would have to unite behind one proposal. Numerous legislative proposals already sit on Capitol Hill, ranging from independent reviews to a federal kill switch to a total moratorium on data centres proposed by Senator Bernie Sanders. These diverging proposals point to a lack of consensus in Congress, but also to an urgency among some lawmakers to take action. Republican Senate Majority Leader John Thune and his Democratic colleague Amy Klobuchar are working on a bill that would force some AI companies to submit to federal oversight. "You don't want to stifle innovation," Thune said on Monday, "but I think you also want to make sure that the more advanced threats can be mitigated." Democratic congresswoman Lori Trahan has spent months working on a different bipartisan proposal in the House that would give the government power to halt the use of an AI model if officials find there is an imminent catastrophic risk. "The federal government needs to put guardrails in place," she told the BBC, warning that the technology would eventually affect every aspect of society. "I don't think we have any more time to wait. This is all hands on deck." Trahan has worked on her proposed legislation, the Frontier Act, jointly with Republican Jay Obernolte, and among other things it would require independent audits of AI research labs. She recognised, nonetheless, that any movement this week was unlikely.

Some leading Republicans, including Speaker Johnson, have often referred to self-regulation when discussing government action and called on companies to take responsibility for AI safety. One of Trump's top AI advisers, David Sacks, told the BBC's US partner CBS News on Monday that some of the existential warnings by insiders had been overblown. "I think this is becoming a panic," he said. "First and foremost, it is on them to make their products safe," Sacks said. "They should tell us how they should do that. If they can't do that, they should step aside for people who can do that." Referring to Anthropic's boss Dario Amodei, who has called for a slowdown in development, he added: "Why are you acting like this is something that you can't control? If you can't control it, then don't do it."

Some experts share the view that, in the short term at least, it is up to the AI companies to regulate themselves responsibly. "You might want an act of Congress, and I would love a congressionally mandated regime that requires safety and testing and bias testing," said Asad Ramzanali, director of AI and Technology Policy at Vanderbilt University. "But absent that action, the companies have autonomy." Others have expressed scepticism at the recent flurry of warnings by AI staffers. Trevor Traina, a tech executive who served as a diplomat during Trump's first term, said: "I don't think the typical citizen has encountered anything scary or really problematic with AI." He added: "We're relying on the tweets of a half-dozen people who are all vying for perceived supremacy. The cynic in me thinks, are they really concerned, or are they trying to prove which model is the most awesome?"

Alexandra Reeve Givens, who previously worked in government and now leads the US non-profit Center for Democracy and Technology, said she did not think meaningful federal AI regulation would pass in the foreseeable future. While the Trump administration has set up an entirely voluntary framework for AI companies to submit their models and tools for government assessment, Givens noted that the framework itself remains a secret. "There's no clear standard to make sure that the approach is even grounded in the rule of law," she said. Appeals from various groups to release the framework have so far gone ignored. The White House approach, combined with a lack of political will and consensus, makes passing laws around AI difficult. "There is no question that federal legislation is challenging in this climate, but Congress also has to grapple with which of the many types of AI risk need to be addressed," Givens said. "It's not that AI is ungovernable, it's that each risk needs a tailored approach."