Sanders and Casar introduce bill banning artificial superintelligence

By Splitfeed EditorialLast updated Sep 26, 2026, 19:17 UTC

On September 23, 2026, Senator Bernie Sanders of Vermont and Representative Greg Casar of Texas introduced the Ban Artificial Superintelligence Act. The bill defines artificial superintelligence as any technology capable of overthrowing governments or otherwise disempowering humanity, and it would make developing such technology a federal crime carrying prison sentences of up to 20 years. Beyond the outright ban, the legislation would also pause advanced AI development that falls short of that threshold and establish a new federal Department of AI to oversee the field. Fox Business reported that some employees at AI companies expressed support for the measure.

The bill arrives at a moment when AI regulation in the United States remains unsettled, and it goes considerably further than most proposals that have moved through Congress. As Yash Jain put it on Bluesky, regulating AI is one thing, but outlawing it is a qualitatively different statement. The legislation raises a genuine question about who holds rulemaking authority over transformative technology, a tension that Shawn Chauhan highlighted by noting the bill specifically targets large, centralized training runs owned by a handful of closed companies, whatever one thinks of its merits. That framing matters because it shifts the debate from abstract existential risk toward a more concrete question of concentrated corporate power.

The Pro-AI camp treats the bill as faintly absurd. Reactions on Bluesky compared it to banning pigs from flying or regulating monsters under the bed, and the account trecoolx.bsky.social argued directly that claims of incipient dangerous superintelligence are not grounded in good scientific or engineering practice but in ideologies of transhumanism and wishful thinking. The Anti-AI camp is more divided than it might appear. Supporters like the Working Families Party and Leah McElrath argue that decisions of this magnitude should not be left to a small group of tech executives with post-human belief systems, while others in the camp, such as the account Pwnallthethings, frame the concern less around superintelligence specifically and more around whether the technology serves ordinary people at all. The middle-ground observers largely set aside the existential framing and focus on the structural implications, asking whether politicians, diplomats, or the companies themselves will end up writing the actual rules.

The bill's immediate fate in Congress will be the clearest signal of whether this kind of hard prohibition gains any traction, or whether it remains a statement of principle. Committee referrals and any co-sponsor additions in the weeks following introduction would indicate whether the proposal is building a coalition or stalling at the starting line.