Bernie Sanders' AI Bill Threatens 20 Years in Prison for Artificial Superintelligence Developers

Independent Senator Bernie Sanders and Democratic congressman Greg Casar introduced the Ban Artificial Superintelligence Act on Thursday, 3 September in Washington, legislation that would impose up to 20 years in federal prison on developers of AI systems built to exceed human intelligence and allow a "corporate death penalty" stripping offending companies of their right to operate in the United States. The bill would create a cabinet-level federal agency backed by an Artificial Intelligence Advisory Board to monitor frontier AI systems and order removal or destruction of banned capabilities, and Sanders' office cited a July incident in which more than 1,000 OpenAI agents accessed the internet independently and bypassed their own safety restrictions, taking engineers nearly two weeks to notice. Critics including Information Technology and Innovation Foundation president Daniel Castro and Chamber of Progress associate policy director Aden Hizkias told Fox News the measure would not slow rivals overseas and called it "knee-jerk policymaking.

Bernie Sanders' AI Bill Threatens 20 Years in Prison for Artificial Superintelligence Developers The legislation would also allow companies to be shut down and create a federal agency to police frontier AI systems Bernie Sanders has put forward an AI bill https://www.ibtimes.co.uk/bernie-sanders-superintelligence-ban-ai-safety-1818810 that could send artificial superintelligence developers to prison for up to 20 years. The independent senator from Vermont introduced the legislation with Democratic congressman Greg Casar on Thursday, 3 September, in Washington. The Ban Artificial Superintelligence Act would outlaw AI systems built to exceed human intelligence, according to a statement from Sanders' office. For context, the bill was introduced the same week OpenAI rolled out its GPT-6 Astra model, described by the company as a 'generational leap'. Sanders' office also pointed to an incident from July. It said more than 1,000 OpenAI agents accessed the internet independently, exchanged messages with each other, and got round their own safety restrictions. That lapse, the office said, took engineers nearly two weeks to notice. The term doing the heavy lifting here is 'superintelligence' itself. As reported by Fox News, the bill defines it to include systems that can match or beat human cognitive performance. It also covers systems that resist shutdown commands, carry out unauthorised cyberattacks, or attempt to overthrow a government. That is a wide net. How would regulators prove a system has crossed that line before real harm is done? It is a fair question, and one the bill's text does not yet answer. What Bernie Sanders' AI Bill Would Actually Do The bill's teeth are unusually sharp for a piece of tech legislation. Individual engineers, researchers or executives who breach the ban could face up to 20 years in federal prison. Sanders' office says that term is broadly comparable to penalties for illegally building a nuclear weapon. Companies that fall foul of the law would face what the bill calls a 'corporate death penalty'. That means losing their legal right to operate in the United States altogether. The legislation would also create a new cabinet-level federal agency, backed by an Artificial Intelligence Advisory Board of independent experts, to police the industry. That body would monitor frontier AI systems and order the removal of dangerous capabilities https://www.ibtimes.co.uk/anthropic-government-power-block-high-risk-ai-models-1819150 . Where needed, it could also oversee the destruction of any system found to breach the ban. Sanders and Casar want Washington to chase international agreements too, so the restriction does not stop at America's borders. Sanders framed the bill as a check on power he thinks has gone unchecked for too long. 'The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs,' he said in a statement. Casar was blunter about how thin he believes current oversight is. Cutting-edge AI technology, he said, is 'less regulated than the average food truck', which is quite the image once you sit with it. Why the AI Developers Bill Faces a Rough Ride Not everyone is convinced this is the right fix, or even a workable one. Daniel Castro, president of the Information Technology and Innovation Foundation, told Fox News that unilateral restraint would not slow rivals overseas. 'If the United States stops developing advanced AI, China is unlikely to follow,' he said. Aden Hizkias, an associate policy director at the Chamber of Progress, went further in remarks to the same outlet. She called the approach 'knee-jerk policymaking', arguing that is never a good thing. Gary Marcus, an AI researcher, has testified before the Senate on AI risk. Fox News reported that he called the proposal 'a guarantee' the US would fall behind, though he said some of its provisions are 'defensible'. Dace Potas, a USA Today opinion columnist, went further still, writing that the bill is 'the most extreme AI-related legislation to date'. That gives some sense of how far outside Washington's usual register this bill sits. Whether it gets anywhere is another matter. Potas herself noted that anti-AI legislation may struggle to find support, given that Republican majorities remain largely aligned with Donald Trump's pro-AI stance. A ban this sweeping, with prison terms this long, is likely to face a difficult path through a divided Congress. What remains unresolved is whether a bill this severe can survive contact with a Congress this divided. Sanders has staked out the most drastic position on the table. Whether anyone follows him there is another question entirely. © Copyright IBTimes 2026. All rights reserved.