Sanders Unveils Bill to Ban Superintelligent AI and Create a New Federal Department

Ban the mind that outruns man,
Build a building, draft a plan.
Twenty years if code should stray—
Pause the future, start today.

Sen. Bernie Sanders and Rep. Greg Casar introduced the Ban Artificial Superintelligence Act on Wednesday. The measure would permanently forbid systems that match or exceed human cognition across domains or that could plan humanity’s disempowerment, pause the most advanced model work until new rules exist, and stand up a cabinet-level Department of Artificial Intelligence.

A Ban Written in Maximums

The bill defines superintelligence broadly and attaches nuclear-style penalties: up to 20 years in prison for people and a “corporate death penalty” for firms that try to evade the pause. A new agency would monitor frontier systems, order dangerous capabilities stripped, and supervise destruction of banned models. An advisory board of experts would sit beside the secretary.

Safety Case Versus Control Case

Sanders says company leaders already admit they do not fully understand the systems they ship. Supporters inside some labs signed on. Critics note that “superintelligence” has no agreed technical definition, that a pause hands advantage to foreign labs, and that a brand-new department is a permanent bureaucracy built on a moving target.

Congress as Product Manager

Advanced systems would need federal approval before deployment once rules exist. The bill also urges international treaties to export the ban. Passage is uncertain. The text arrives as models grow more capable and as other capitals race rather than halt.

The Practical Question

If the threat is real, narrow, testable limits beat a new ministry. If the threat is speculative, criminalizing research that cannot be cleanly defined invites selective enforcement and capital flight. That argument starts now.