Congress now has a bill that would make building a machine smarter than humanity a federal crime. On Wednesday, September 23, Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) introduced the Ban Artificial Superintelligence Act. It would permanently prohibit artificial superintelligence, halt the most advanced AI development until a new federal regulator writes safety rules, and create a Cabinet-level Department of Artificial Intelligence to enforce both. "When you are racing towards a cliff, you don't just ease up on the gas pedal. You hit the brakes," Sanders said in a statement announcing the bill. "When the future of humanity is at stake, we cannot let a handful of Big Tech CEOs write their own rules."
What the Bill Would Do
The bill defines artificial superintelligence as an AI that exceeds human cognitive performance and capabilities across most domains, or one with enough capability to destroy or disempower humanity, including by overthrowing the federal government. No person or company could develop or deploy such a system. Advanced AI development would also stop until the new department is running and has set up rules and a model review process. After that, according to Roll Call, companies would have to submit pre-development plans to the department and get final approval before any public deployment.
The penalties are deliberately modeled on nuclear weapons law. Companies that violate or try to get around the pause or the ban would face what the lawmakers call the "corporate death penalty," and individuals could get up to 20 years in prison. The new department, advised by a board of scientists, would monitor frontier systems and oversee the removal of dangerous capabilities such as subverting shutdown commands. Another provision makes it US policy to pursue international agreements, allied coordination, and export controls to stop superintelligence from being built anywhere.
Casar, who chairs the Congressional Progressive Caucus, said the bill would not freeze all AI. He pointed to its ban on systems showing "superintelligence precursor characteristics." These include the ability to help design nuclear, chemical, or biological weapons, to deceive humans, or to modify their own functions. "Experts are warning that AI superintelligence, wielded by the wrong humans or by rogue AI, could kill countless numbers of people. But Donald Trump says he wants to encourage it. Something has to change," Casar said. He also told reporters that voluntary industry commitments based on each company's own risk tolerance are "not how this works in any other industry."
Support From Inside the Labs
Some of the most notable support comes from people who work at frontier labs. The Associated Press, which got the bill first, reported that several employees at leading AI companies signed a statement backing it. Juan Felipe Cerón Uribe, a researcher on OpenAI's Safety Systems team, said "superintelligence could either go extremely right or extremely wrong" and that "we shouldn't be playing such games." Swante Scholz, a Google DeepMind software engineer who said he was not speaking for his employer, warned that on the current path the "most likely outcome is an existential catastrophe for humanity."
Sanders was blunter in his AP interview: "It doesn't take a genius to say, 'slow it down.'"
The Case Against
Critics have several objections. The most common is geopolitical. The bill arrived during Chinese President Xi Jinping's state visit to Washington. A day earlier, President Trump told the UN General Assembly that he opposed a "globalist scheme to control" AI, and he framed the technology as a race. "Whoever wins superintelligence wins," Trump said, adding that the US is "leading now over China by a lot." Opponents argue that a US-only pause would give that lead to Beijing. Casar acknowledged that some in the tech industry have pushed back on a mandatory domestic pause without an international agreement in place.
Other groups that share the bill's concerns still question its approach. Brendan Steinhauser, CEO of the Alliance for Secure AI, said Sanders and Casar "deserve credit for taking superintelligence seriously" but did not endorse the ban. "The Alliance believes that a more measured approach is necessary to get this right," he said.
A third critique comes from pro-competition voices. They argue that a licensing-and-approval regime run by a powerful new agency could lock in today's leading labs, which can best afford compliance, and shut out smaller rivals and open-source developers. That concern carries more weight this week because a separate antitrust class action alleges that major labs coordinated a slowdown.
Why It Matters
The Ban Artificial Superintelligence Act is not expected to pass. The AP noted that it faces long odds in a Republican-controlled Congress that has struggled to agree on much milder AI rules. Its importance is in setting a marker. It is the most aggressive federal AI proposal introduced so far, and it moves the progressive position from transparency and liability to outright prohibition backed by prison time. It also sets progressives against a White House that rejects new regulation and House Republicans pushing federal preemption of state AI laws.
The timing matters too. Industry leaders, including Anthropic CEO Dario Amodei, have recently called publicly for pacing frontier development, and some lab employees are now endorsing statutory limits. That makes it harder to dismiss the bill as fringe. Supporters say that if companies admit the risks are real, voluntary restraint is not enough. Skeptics say a hard legal ban risks slowing useful research, ceding ground to rivals, and concentrating power in the firms already at the frontier. Both arguments will shape AI policy fights ahead of November's midterms.
What to Watch
The bill is not yet numbered, so its committee referrals and any cosponsors beyond Sanders and Casar are the first test of support among Democrats. Watch whether AI safety becomes a concrete item in the Trump-Xi talks this week, since Sanders has urged the president to negotiate a joint pause and superintelligence ban with China. Also watch how the frontier labs respond now that their own employees have endorsed the bill.
“When you are racing towards a cliff, you don't just ease up on the gas pedal. You hit the brakes.”— Bernie Sanders, U.S. Senator (I-Vt.)