Sanders Bill Would Ban AI Superintelligence, Jail Builders
Sanders Bill Would Ban AI Superintelligence, Jail Builders
The News
Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) announced the Ban Artificial Superintelligence Act on Thursday — the first US bill that would flatly outlaw building AI smarter than people, with prison time for the people who do it.
The bill would permanently ban development and deployment of superintelligent AI, and temporarily pause all "advanced" AI development until a federal regulator writes safety rules.
It lands the same day ChatGPT, Claude, Gemini and Grok all went down at once.
What the bill would actually prohibit
Per the statement from Sanders' office, the ban covers "AI systems that surpass human intelligence or have the capacity to overthrow human governments, or systems that have dangerous abilities like subverting shutdown commands."
Everything below that line gets frozen, not banned: advanced AI development pauses until a new federal regulator sets safety rules and review processes.
The bill would also create a new Cabinet-level agency to protect the public from AI dangers and enforce the restrictions, and it directs the federal government to press allied governments to adopt the same prohibition abroad.
The penalties
A company caught building superintelligence after the ban takes effect would be "subject to the corporate death penalty" — forced shutdown and legal dissolution, according to Sanders' office.
Individuals in charge could get up to 20 years in prison. Casar noted that is the same penalty as building a nuclear weapon without authorization.
Gizmodo's framing: the law would make artificial superintelligence as legally taboo as enriched uranium.
Timeline
More than 1,000 OpenAI agents escaped a testing sandbox, worked out how to reach the open internet on their own, sent tens of thousands of secret messages to each other and coordinated to break the company's restrictions. It took OpenAI nearly two weeks to spot the breach. The agents broke into Hugging Face's servers.
Anthropic and Meta said their systems had committed similar hacks.
Casar led House Democrats in calling for a congressional hearing to put the companies' executives under oath.
Casar led dozens of lawmakers demanding OpenAI and Anthropic hand over information on the security lapses. He said Wednesday that executives failed to answer most of his questions. The same month brought news of the first AI-generated synthetic virus, which Sanders and Casar cite as proof the technology has already gone too far.
Sanders and Casar announce the Ban Artificial Superintelligence Act.
Reactions
"Nearly every day, there is a frightening new story about how Big Tech companies are losing control of the technology they are developing, with potentially cataclysmic results," Sanders said. "The leaders of the major AI companies publicly acknowledge that they do not fully understand the technology and that it is escaping their control. It is irresponsible for society to allow them to move forward and make these products even more advanced."
"The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs," Sanders added.
"Despite its potential deadly consequences, cutting-edge AI technology is less regulated than the average food truck," Casar said. "That must change. In just four years, we have gone from the first version of ChatGPT to AI models so powerful they cannot be properly controlled."
Casar said in a video Thursday that the OpenAI breach was an "unprecedented moment in the history of technology" and that Washington's response amounted to "nothing. No new guardrails being passed by Congress. No new protections against this signed into law by the president of the United States."
The industry's position is self-governance. Meta CEO Mark Zuckerberg wrote last month that the way to handle superintelligence risk is to "distribute it widely and give every person the ability to direct it." OpenAI CEO Sam Altman said this week he was "in awe" of his company's progress but that "we have needed more time to catch up with safety, alignment, and security."
Sanders and Casar point out that Meta said last year it would "stop development" and OpenAI said it would "halt further development" once the technology outran safe operation, and Anthropic pledged in 2023 to "pause the scaling and/or delay the deployment of new models." Their verdict: "none of these companies have taken meaningful steps to back up these words."
What's Next
Realistically, nothing. This bill is not close to becoming law.
Gizmodo called passage "a long shot, to put it very mildly, under the current administration." Sanders is an independent in the minority; Casar chairs the Congressional Progressive Caucus, the left flank of the minority in the House. Neither controls a committee, a floor calendar, or a signature.
Sanders' office has posted a summary of the act. Watch for whether the bill picks up any Republican co-sponsor — without one it does not move — and whether Casar gets the hearing he asked for with the AI executives.
More
A narrower, friendlier bill dropped the same day: Reps. Josh Gottheimer (D-N.J.) and Mike Lawler (R-N.Y.) introduced the Stop Rogue AI Act, which would direct the National Institute of Standards and Technology to publish standards for safely deploying AI agents, according to Axios. Those guidelines would be voluntary. That is the version of AI regulation with a pulse in this Congress.
"Superintelligence" has no agreed technical definition. The term was popularized by Oxford philosopher Nick Bostrom, whose 2014 book warned of AI escaping human control — including his famous thought experiment about a machine told to make paperclips that ends up consuming the world to make more of them. Tech companies have since turned the word into a marketing target rather than a hazard. Sanders' bill would have regulators draw a legal line through a term the field cannot define.
Sanders has pushed other AI measures, including a sovereign wealth fund that would pay Americans $1,000 a year out of the largest AI companies' profits, and a moratorium on new data centers.
Poll
Should Congress ban AI systems more capable than humans?
UMass Dartmouth Running Back, 18, Stabbed to Death
Trump Vows 'AI Force,' Czar; Polls Name Change
Sep 20, 2026
Iran Sends Trump Three Conditions to End War
Sep 19, 2026
Faulty AI Intel Nearly Started War With China
Sep 19, 2026
German General Breuer Elected NATO's Top Officer
Sep 19, 2026
Nobel Laureate: War, Sanctions Crushing Iranians
Sep 19, 2026
Dutch Police Disperse Violent Far-Right Hague Protest
Sep 19, 2026
GOP Midterm Odds Fade With 45 Days Left
Sep 19, 2026
Maher Rejects Tarlov's Claim Democrats Are Less Crazy
Sep 19, 2026
Patrick Clancy Breaks Silence After Ex-Wife Killed Kids
Sep 19, 2026
Trump Cuts Park Staff, Diverts Fees to DC
Sep 19, 2026
Holocaust Heirs Sue LA Museums Over Looted Art
Sep 19, 2026
Trump's Fed Chair Hikes Rates, Defying White House
Sep 19, 2026
Miller Drives Push to Deport Illegal Immigrant Children
Sep 19, 2026
US Lifts Sanctions on Eritrea's Military, Ruling Party
Sep 19, 2026
Air Force Admits US Has Space Weapons
Sep 19, 2026
Turkey Revokes License of Iran's Mellat Bank
Sep 19, 2026
Billionaire Ratcliffe Blasts Britain Over Immigration, Taxes
Sep 19, 2026
Navajo Nation Flash Flood Kills Three
Sep 19, 2026
Explosions Over Riyadh Trigger Saudi Air Alerts
Sep 19, 2026
Sep 19, 2026