Bernie Sanders calls for outlawing 'superintelligence,' with violators facing jail time

Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) have put forward a new bill that would make it illegal for anyone to build artificial superintelligence — a form of AI the legislation characterizes as capable of "the destruction or disempowerment of humanity," potentially including the overthrow of the US government. Under the proposed Ban Artificial Superintelligence Act, executives leading AI companies could face prison sentences of up to 20 years for breaching the ban.

What the Bill Covers

The legislation goes beyond targeting only the most powerful class of AI systems. In addition to prohibiting artificial superintelligence outright, the bill would impose a pause on developing advanced AI systems more broadly, according to reporting by The Verge.

The term "superintelligence" refers to hypothetical AI that would surpass human intelligence across virtually all domains. The bill's framing reflects lawmakers' concerns that such systems, once created, could not be reliably controlled — hence the inclusion of catastrophic outcomes like government overthrow in its definition of potential harms.

Political Context

The proposal arrives as AI terminology has become a political flashpoint. President Donald Trump has said the United States will use "superintelligence" in official documents in place of the term "AI," giving added weight to the language the legislation centers on.

The bill also lands in the middle of an ongoing debate in Congress over how — or whether — to regulate advanced AI. While industry leaders continue to race toward more capable models, proposals to slow or halt frontier development have so far gained limited traction in Washington.

What Happens Next

The legislation is only a proposal at this stage and faces steep odds in a Congress where most AI-related measures have stalled. Whether it gains committee attention or co-sponsors will signal how much appetite exists on Capitol Hill for hard limits on frontier AI development rather than narrower safety rules.

For now, the bill stands as one of the most sweeping AI restrictions ever introduced at the federal level, and its progress will be worth tracking as the superintelligence debate intensifies.