Sanders and Casar Bill Would Permanently Ban Superintelligent AI

Senator Bernie Sanders and Representative Greg Casar introduced a bill on September 23 that would permanently ban the development and use of artificial superintelligence in the United States.

The legislation, put forward by the Vermont independent senator and the Texas Democratic representative, represents one of the most aggressive regulatory positions taken to date in the U.S. Congress on frontier AI development. Rather than imposing licensing regimes, safety thresholds, or interim moratoriums, the bill targets superintelligence itself — a permanent prohibition on building and deploying it.

The proposal would also create a new federal AI agency, giving the government a dedicated bureaucratic structure to enforce the ban and oversee artificial intelligence development more broadly.

Why it matters

The bill lands amid an intensifying debate in Washington over how — and whether — to regulate frontier AI systems. Major AI developers including OpenAI, Google, Anthropic, and Meta are racing toward systems they describe as increasingly general and capable, while some researchers and executives have openly discussed superintelligence as a plausible engineering goal. A permanent statutory ban would place U.S. law in direct conflict with the stated ambitions of the country's most heavily funded technology companies.

Defining "artificial superintelligence" in legally enforceable terms will likely prove to be the bill's central challenge. Any permanent ban hinges on a statutory definition robust enough to distinguish prohibited systems from the powerful but narrower models that dominate today's market — a distinction that remains contested even among AI researchers.

The proposed federal AI agency also signals a structural shift in how at least some lawmakers want AI governed: not through existing regulators like the FTC or NIST frameworks, but through a purpose-built institution with its own mandate.

Long odds, real signal

As a bill introduced by two progressive lawmakers, the proposal faces steep odds in a Congress where AI legislation has stalled along partisan lines and where competing proposals — from light-touch innovation frameworks to licensing regimes — are still fighting for traction.

But the legislation establishes a marker on the far end of the regulatory spectrum. If debate over AI safety and superintelligence advances in the current session, other proposals will be measured against a position that treats superintelligence not as a risk to manage, but as a capability to outlaw entirely.

The bill's fate in committee will indicate whether outright prohibition moves from the fringe of AI policy debate toward its center.