Sanders Unveils Radical Bill to Ban Artificial Superintelligence and Freeze Advanced AI

On September 23, 2026, Senator Bernie Sanders and Representative Greg Casar introduced a sweeping bill that would ban the development and deployment of “artificial superintelligence” and pause the most advanced AI projects in the United States. The proposal defines artificial superintelligence as any system that surpasses human cognitive performance across most domains, or that has the capability to destroy or disempower humanity, including by toppling the federal government.

The legislation would create a new federal agency with authority to license and strictly oversee frontier‑level AI development. Systems above a certain capability threshold — particularly those with potential military uses or the ability to disrupt critical infrastructure — would be prohibited outright. Beyond U.S. borders, the bill makes it official policy to work through export controls, alliances, and international agreements to prevent the emergence of superintelligence anywhere in the world.

Politically, the measure clashes head‑on with the Trump administration’s “full‑speed‑ahead” stance on AI and faces long odds in the current Congress. Still, it reflects growing unease among researchers and policymakers about existential risks from runaway AI systems. The debate over this bill is likely to sharpen a core question for global AI policy: when does prudent safety planning turn into an attempt to halt technological progress altogether?