Senator Bernie Sanders (I-Vt.) and Representative Greg Casar (D-Texas) on September 23 filed matching bills in the Senate and House that would permanently prohibit the development of superintelligent AI and create a new cabinet-level AI agency to license, audit and, if necessary, shut down frontier developers.
The Ban And The Temporary Pause
The legislation defines superintelligence as an AI system that "meaningfully exceeds" human performance across substantially all cognitive tasks. It would impose a temporary pause on training runs above a compute threshold to be set by the new agency, and a permanent prohibition on deploying systems that meet the superintelligence definition.
A New Department Of AI
The bill establishes a cabinet-level Department of AI headed by a Senate-confirmed secretary. The department would maintain licensing regimes for frontier developers, run pre-deployment safety evaluations, and hold statutory authority to compel model shutdowns. Companies operating over the training-compute threshold — a group that today would include OpenAI, Anthropic, Google DeepMind, Meta, xAI and Microsoft's in-house teams — would need continuous licenses.
Where It Stands
The Sanders-Casar bill is unlikely to reach the floor of a Republican-controlled Congress this year, but it lands as United Nations Security Council members held their first formal briefing on AI safety this week and as California Governor Gavin Newsom signed a separate executive order requiring an emergency shutdown mechanism for frontier models operating in the state.
Related coverage: Newsom's California AI kill-switch EO, Anthropic ships Claude Opus 5.5, and OpenAI's GPT-6 Sol and Luna price cut.
Reporting based on coverage from sanders.senate.gov, Reuters and industry announcements on September 23, 2026.