Rep. Greg Casar, D-Texas, is calling for the United States government to end what he described as an artificial intelligence (AI) arms race. He said the federal government should set limits on frontier development rather than leave those decisions to technology companies.
Speaking today at the POLITICO Decoded Summit in Washington, D.C., Casar called for a pause on frontier AI development while Congress establishes federal safeguards and the United States pursues an international agreement with China.
“The AI arms race is not a race that you win,” Casar said. “We’re like in the car barreling towards a cliff, and if you win the race, you fall off the edge of the cliff.”
“I think that the United States should be leading in AI and technological development, but also be leading the world in AI diplomacy, because this is not an arms race that you win,” he said. “It’s an arms race that we need to end.”
Casar stressed that he is not calling for a halt to all AI development. Instead, he said companies should pause development of the most advanced systems while the federal government creates an agency to oversee AI safety.
The lawmaker said that pause should be paired with an “enforceable and verifiable treaty with China” intended to prevent dangerous AI systems from being developed elsewhere.
“By putting guardrails in place now to protect Americans, we show China that we’re serious about getting to a deal,” Casar said, adding, “Chinese models currently get trained on American models. So, by putting a pause on the most extreme American frontier development, we actually slow China down as well.”
That approach is at the heart of forthcoming legislation Casar is pursuing with Sen. Bernie Sanders, I-Vt.
The Ban Artificial Superintelligence Act would permanently prohibit the development and deployment of superintelligent AI – AI that humans cannot control. It would also temporarily pause advanced AI development until a federal regulator establishes safety rules.
Casar said Wednesday that formal bill text will be released “very shortly.”
The proposal would also establish a cabinet-level agency to oversee advanced AI development. Casar rejected relying solely on company-funded independent auditors, comparing the potential risks of advanced AI to those posed by nuclear technology.
“If you’ve got the CEOs of the big companies, like Sam Altman and Dario Amodei, signing onto statements saying this could become as dangerous as nuclear weapons, then you should actually have a federal agency just like we had around nuclear weapons,” Casar said.
The Texas Democrat said top AI company CEOs should also testify before Congress under oath to provide greater visibility into safety incidents involving advanced models.
Casar also called on Congress to “wake up and put in place protections” to address other AI risks, including mass unemployment, mass surveillance, and mass destruction.
Casar’s remarks come as the debate over AI safety has intensified in recent days. President Donald Trump on Sept. 14 rejected calls for additional federal AI regulation, arguing that the government already has sufficient authority over AI companies and that further restrictions could undermine U.S. competitiveness.
In a series of posts on Truth Social, Trump said the only guardrails AI needs are those provided by a “STRONG AND SMART” president.
Trump’s comments followed calls from several leading AI executives to slow the development of increasingly powerful models.
Anthropic CEO Dario Amodei called for developers to pace advances in AI capabilities, a position also backed by OpenAI CEO Sam Altman and xAI CEO Elon Musk. Trump dismissed warnings about AI posing catastrophic risks as a “HOAX” and said restrictions would benefit China as the two countries compete for AI leadership.
Rep. Brett Guthrie, R-Ky., who also spoke at today’s Politico event, took a similar position to Trump, saying, “The idea that we’re going to limit America [in frontier AI] and expect China to say, ‘Oh, that was very altruistic. We’ll do the same thing,’ I think, is extremely naive and extremely dangerous.”
Guthrie, who chairs the House Energy and Commerce Committee, added that a bipartisan proposal to put guardrails around AI may not come until 2027.
“I think a bill of that magnitude needs work,” Guthrie said, adding, “I wouldn’t want to do something in a lame duck session to do it quickly and not get it right.”