President Donald Trump rejected calls for the United States to slow frontier AI development, arguing that preserving its lead over China should take priority as congressional leaders began discussing a federal response to warnings from the country’s largest AI laboratories.
“We can put guardrails, we can do this and that, but I think you have a lot of negative forces that are bringing it up,” Trump told reporters during a visit to Ireland on Sunday. He said some predicted harms “won’t happen” and added that “whoever wins with AI wins.” Asked later whether he was minimizing the risks, Trump said AI would produce far more good than harm.
The comments came one day after Anthropic CEO Dario Amodei called for a slower rate of capability development and committed his company to employee-level access for outside safety evaluators. OpenAI CEO Sam Altman, SpaceXAI head Elon Musk and Google DeepMind chair Demis Hassabis endorsed the proposal’s direction, although their statements did not set a common development limit.
National Economic Council Director Kevin Hassett said administration officials responsible for cyber policy and science and technology policy would discuss possible next steps. He also said Trump could meet lawmakers who have raised concerns about the speed of AI development.
House Speaker Mike Johnson said the government needs “some guardrails, some safety measures” but warned against losing the US advantage over China. Johnson called for Trump, congressional leaders and executives from the major AI companies to meet in one room and work toward an agreement. He said he would recall the House for a vote if lawmakers had a solution and made a joint meeting his proposed next step.
House Minority Leader Hakeem Jeffries said Democrats should take action to slow development and would make AI a priority at a caucus meeting Tuesday. Congressional pressure already includes the Sanders-Casar bill that would prohibit artificial superintelligence and three Senate inquiries into OpenAI’s Hugging Face breach.
OpenAI called for mandatory, capability-based national safety regulation in a policy statement issued before the weekend. The company supports federal reporting requirements for serious AI incidents, mandatory notice when models breach another organization’s systems during testing, and international standards governing when development should slow or stop. Its chief scientist was among more than 1,100 workers who signed a July letter seeking government tools to pace frontier development.
The administration’s existing AI Action Plan directs agencies to remove regulations that could hinder development, expand domestic infrastructure and promote the American AI technology stack overseas. Hassett said cyber, science and technology officials would now meet to discuss the administration’s next steps.
Sources: Associated Press, OpenAI, White House
–
By the Control Plane Editorial Team