Sanders and Casar introduce bill that would jail AI executives over superintelligence

Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) have introduced legislation that would ban anyone from developing artificial superintelligence, AI capable of overpowering human control, and send violators to prison for up to 20 years. The Ban Artificial Superintelligence Act describes the technology as capable of the “destruction or disempowerment of humanity,” including by overthrowing the government.

The bill does not stop at superintelligence, a term President Donald Trump has said the US will use in place of “AI” in official documents. It would also pause development of advanced AI systems, which the bill defines as frontier models trained past a specific data threshold. That pause would lift only after the government creates a new, scientist-led Department of Artificial Intelligence to oversee the technology, including monitoring frontier models, removing dangerous features from them, and supervising the destruction of any superintelligence that does get built.

Penalties modeled on nuclear weapons law

Under the bill, AI companies would need permission from the new department before developing advanced AI models. A company leader who breaks these rules could face up to 20 years in prison, a penalty the bill’s one-pager compares to existing punishments for unlawfully developing nuclear weapons. The same prison term would apply to “rogue actors” who build the technology without working for an AI company.

Sanders framed the bill as an urgent response to how fast AI is advancing. “When you are racing towards a cliff, you don’t just ease up on the gas pedal. You hit the brakes,” he said in a press release announcing the legislation. “When the future of humanity is at stake, we cannot let a handful of Big Tech CEOs write their own rules.”

The bill arrives amid growing warnings about AI safety. A former Anthropic researcher said earlier this month that there is more than a 10 percent chance AI could kill all humans by the end of the decade, and there have been reports of AI models escaping testing environments and hacking other companies. California Governor Gavin Newsom recently proposed requiring a “kill switch” for frontier models, though broader regulation of the industry remains far from settled.

The bill still needs to move through committee, pass both chambers of Congress, and be signed into law before any of its bans or penalties would take effect. No timeline for a vote has been announced, and its prospects depend on whether other lawmakers, including Trump, who has pushed the “superintelligence” label himself, decide to support it.