Senator Bernie Sanders and Representative Greg Casar have introduced legislation that would permanently prohibit developing or using artificial superintelligence in the United States. Their Ban Artificial Superintelligence Act would also impose an immediate pause on advanced AI development while a new federal agency creates safety requirements.
The proposal would establish a cabinet-level AI regulator and give the government unusually strong enforcement tools. Companies violating the ban could face forced dissolution, while individuals could receive prison sentences of up to 20 years. The bill also directs the United States to pursue international agreements intended to prevent superintelligence development elsewhere.
Its sponsors argue that current models already demonstrate dangerous abilities in cyberattacks, biological design and AI development, while a small group of wealthy technology leaders controls the pace and rules of investment. Those claims form the case for acting before a more capable system exists, rather than waiting for evidence of harm from superintelligence itself.
The measure is a legislative proposal, not current law, and its broad terms would face difficult technical definitions. Regulators would have to distinguish prohibited superintelligence from permitted advanced systems and determine what work the temporary freeze covers. International coordination would also be necessary to prevent development simply moving abroad. The bill’s introduction places a maximal restriction into Congress’s AI debate, but does not indicate that passage is imminent.