What happened
In the wake of the 3 November 2026 US midterms, Rep. Yassamin Ansari (D-AZ) publicly urged Democrats to make 'a pause on advanced AI development, regulation of this out of control industry, and a ban on artificial superintelligence' a priority from 'day 1' of the incoming Democratic-controlled Congress (4-6 November 2026). Ansari is an original co-sponsor of the Ban Artificial Superintelligence Act of 2026 — introduced as H.R. 10538 by Rep. Greg Casar on 24 September and S. 5493 by Sen. Bernie Sanders on 23 September. The bills would create a Department of Artificial Intelligence to oversee and enforce safety requirements, impose a pause on covered advanced-AI training/modification/deployment until the Department is staffed and rules are set, ban 'artificial superintelligence' and systems that could help create it, and impose fines plus up to 20 years' imprisonment for reckless violations. The House bill is referred to Oversight and Government Reform; the Senate bill to Commerce, Science and Transportation.
Why it matters
This is a concrete legislative vehicle for a federal pause-and-ban regime on frontier AI, and its political salience has jumped now that Democrats will hold both chambers. Although introduced in late September (outside this window), the in-window development is the formal positioning of the ban as a top legislative priority for the new Congress, materially raising the prospect of a mandatory, non-voluntary federal frontier-AI oversight and preclearance regime — a sharp contrast with the Trump administration's voluntary EO 14409 approach.
Action needed
Frontier-AI labs and their investors should track committee action on H.R.10538/S.5493 in the 120th Congress, and begin assessing exposure to a compute-threshold-based pause/ban and a possible new federal AI regulator, including how a 10^25 FLOP threshold would be treated by their training roadmap.