The Ban Artificial Superintelligence Act is a federal bill that would permanently prohibit the development of superintelligent AI in the United States and pause work on the most advanced systems until new safety rules exist. It has not become law, and it faces long odds.
Senator Bernie Sanders, an independent from Vermont, and Representative Greg Casar, a Democrat from Texas, unveiled it on September 23, 2026. It is back in the news this week, so here is what it would actually do.
In this article
What the Ban Artificial Superintelligence Act would do
According to WCAX, the bill has four main parts:
- **A permanent ban on artificial superintelligence,** defined as a system that surpasses human cognitive ability or can plan and carry out humanity’s destruction or disempowerment.
- **A temporary halt** on development of the most advanced AI systems until a new Department of Artificial Intelligence sets federal safety rules.
- **Federal approval** before advanced AI systems can be deployed.
- **Penalties,** including up to 20 years in prison in some cases.
Tom’s Guide adds that the bill would create a cabinet-level department with authority to monitor advanced AI development, inspect models before release and order dangerous systems to be disconnected from the internet, modified or rendered inoperative. It also calls for international agreements to stop superintelligence being built elsewhere.
Who is backing the Sanders AI bill
Sanders told the Associated Press, “It doesn’t take a genius to say, ‘slow it down.'” WCAX reports that several current AI company employees and outside experts endorsed the bill, including an OpenAI safety researcher, Juan Felipe Cerón Uribe, who said superintelligence could go “extremely right or extremely wrong.” A Google DeepMind engineer, Swante Scholz, spoke for himself, not his employer, in support.
Why the odds are long
The bill sits in committee. WCAX says it faces long odds in a Republican-controlled Congress, and Tom’s Guide notes that the Trump administration has emphasized AI development and American competitiveness. Sanders has said President Trump has “very little understanding” of the issue and is mostly concerned about the economic effect of slowing AI development.
Critics of strict limits make a related argument: tight approval rules could slow US companies while rivals elsewhere keep building, and an international agreement would be hard to negotiate. Supporters answer that the risk of losing control of a system more capable than humans is worth the delay. The bill itself leaves that question to Congress.
What it means for ChatGPT, Claude and Gemini today
Nothing changes right now. Tom’s Guide stresses the bill has not passed, and that there are no new restrictions on consumer AI services. Even if it became law, it would not automatically ban existing chatbots; its limits target superintelligence, dangerous capabilities and advanced systems that meet certain technical thresholds. The likely effect would be slower or more reviewed releases of future, more capable models.
Why lawmakers are focused on superintelligence
Tom’s Guide says supporters point to warnings from AI researchers and industry leaders about increasingly autonomous systems, and to the concern that companies could build systems whose capabilities they do not fully understand or control. Sanders put it in a September 23 statement: when you are racing toward a cliff, you hit the brakes rather than easing off the gas.
The bill arrives at a tense moment for AI policy. WCAX notes that the Trump administration has resisted calls for more oversight, which puts it at odds with tech leaders such as Anthropic’s Dario Amodei, OpenAI’s Sam Altman and xAI’s Elon Musk, who have each called for more guardrails. Sanders also proposed a sovereign wealth fund in June, financed by a one-time tax on the stock of the largest AI companies.
What to watch next
- **Committee action:** The bill remains in committee, so watch for hearings or markups.
- **The administration’s response:** Its focus on AI competitiveness is the main political obstacle.
- **Industry reaction:** Both supporters and opponents inside AI companies are likely to keep weighing in.
- **International talks:** The bill calls for agreements, which would be a major hurdle.
For more on how AI policy and products are moving, see our AI coverage and our report on ChatGPT Images 2.6.
Ban Artificial Superintelligence Act FAQ
**Is the bill law?** No. It was introduced in September and remains in committee, per Tom’s Guide.
**Does it ban ChatGPT?** No. It targets superintelligence and advanced systems that meet certain requirements, not today’s chatbots.
**What are the penalties?** Fines and prison terms of up to 20 years for certain people who recklessly violate the law, and companies could lose authorization to operate.
Sources: WCAX, Tom’s Guide