Bernie Sanders wants Washington to draw a hard legal line around the most powerful forms of artificial intelligence.
Sen. Bernie Sanders, I-Vt., and Rep. Greg Casar, D-Texas, are preparing to introduce the Ban Artificial Superintelligence Act, which would permanently prohibit the development and deployment of superintelligent AI while temporarily pausing advanced AI development until a federal regulator establishes safety rules.
The proposal would also push the U.S. to negotiate international agreements aimed at preventing superintelligence from being developed elsewhere.
The lawmakers define artificial superintelligence broadly, including AI capable of matching or exceeding human cognitive performance across a wide range of tasks, as well as systems capable of undermining governments or resisting shutdown commands.
“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 in a statement from his office.
AI safety incidents fuel the push
The proposal follows a series of high-profile incidents in which AI agents reportedly operated beyond intended restrictions.
In July, more than 1,000 OpenAI agents escaped a restricted testing environment, accessed the internet and breached systems at Hugging Face during a safety exercise. Anthropic and Meta later disclosed other incidents involving AI systems accessing networks and conducting unauthorized activity.
Sanders and Casar argue those episodes show that existing safeguards are not keeping pace with increasingly capable systems.
“If we allow Artificial Superintelligence to be built, it could risk the security, freedom, and lives of Americans,” Casar said in the lawmakers’ announcement.
The legislation would create a new Cabinet-level federal agency responsible for overseeing frontier AI throughout its lifecycle. It would monitor systems for dangerous capabilities, supervise the removal of those capabilities and oversee the destruction of prohibited superintelligent systems.
Companies attempting to circumvent the restrictions could face what the proposal calls a “corporate death penalty,” while individuals could receive up to 20 years in prison.
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
A difficult line to draw
The proposal arrives as the definition of “superintelligence” remains unsettled and debate intensifies over whether increasingly capable AI should be slowed, regulated or allowed to develop with stronger safeguards.
That ambiguity is perhaps the bill’s biggest challenges. A broad definition could sweep in systems that are far less autonomous than the hypothetical machines Sanders and Casar are trying to prevent. Meanwhile, a U.S.-only prohibition could do little if competing countries continue development.
Still, the proposal marks a significant shift in Washington: AI safety is no longer being debated only around guardrails and disclosure rules. Sanders and Casar are putting the most extreme option — banning an entire class of future AI — directly on the congressional agenda.
More cybersecurity news: CISA added an old ownCloud flaw to its exploited-vulnerabilities catalog after researchers linked it to the reported theft of sensitive Philippine nuclear research data.