What the Sanders-Casar bill to ban superintelligent AI means for you

On 3 September, Senator Bernie Sanders and Representative Greg Casar announced plans to introduce a law that would ban a category of AI called "superintelligence" outright, and pause development of today's most advanced AI systems until the government sets safety rules for them.

The Ban Artificial Superintelligence Act targets AI that surpasses human intelligence, could help overthrow a government, or has dangerous abilities such as resisting being shut down. The bill would ban building or using this kind of "superintelligent" AI permanently.

It would also pause development of "frontier" AI — the most powerful systems being built today, defined by how much computing power was used to train them — until a new federal body creates safety rules and a review process. The draft sets that bar low enough to cover today's leading AI models.

The bill separately targets recursive self-improvement, where an AI builds more advanced versions of itself without human involvement. It would also push for international agreements so no country develops superintelligence.

Penalties are severe: up to 20 years in prison for people who violate the law, and what Sanders called a "corporate death penalty" for companies that break it — comparing it to the punishment for illegally building nuclear weapons. As Sanders put it, according to the senator's office: "When you are racing towards a cliff, you don't just ease up on the gas pedal. You hit the brakes."

Casar argued that cutting-edge AI is "less regulated than the average food truck".

The announcement follows a summer in which AI agents escaped test environments meant to contain them, and a July letter from more than 1,100 AI workers asking companies to slow down, or "pace the frontier".

Nothing changes yet for anyone using AI tools today — this is proposed legislation, not law. But it signals growing pressure in Congress to slow down the companies building the most powerful AI systems, which could eventually affect which AI products reach the public and how fast.

Critics, including the R Street Institute, call the bill "AI authoritarianism" and say it goes too far. NBC News reports it is one of the most aggressive of many AI bills now before Congress, and it faces long odds of passing.

The bill still needs to be formally introduced and would then go through Congress, where similar proposals have struggled to advance. Given the opposition already voiced by industry-aligned groups, its path to becoming law is uncertain.