Cruz-Klobuchar-Thune AI bill nears introduction but faces Senate opposition

A bipartisan Senate AI safety bill from Sens. Klobuchar, Thune, and Cruz may be introduced as early as the week of September 15, with hopes of passing before 2027, but internal disagreements are putting its path in doubt. Ranking member Sen. Maria Cantwell is pushing back over the bill's safety testing provisions, according to Nextgov. Klobuchar has said the bill would require developers to work with government experts to verify and test models, and that dangerous models, including those that evade developer control, should not be released. Transformer News separately reported that a source described the bill as containing no mandatory safety requirements, only a voluntary self-certification regime with no independent evaluations required. Politico reported the bill's prospects remain unclear. Running alongside is the Sanders and Casar Ban Artificial Superintelligence Act, which would halt advanced AI development pending a new cabinet-level regulatory agency and impose penalties including up to 20 years in prison. A new Democratic PAC, Tomorrow Together, launched with over $10 million to push the Democratic Party on AI policy ahead of midterms. Within the Republican Party, a two-day convention in Dallas exposed divisions, with one speaker using nativist framing to oppose AI and Palantir co-founder Joe Lonsdale invoking American frontier technology as a counter-argument.