Both of California’s appropriations committees convene today, Aug. 13, for simultaneous suspense-file hearings on roughly 30 AI bills, according to the Transparency Coalition’s Aug. 7 tracker. It’s the single procedural chokepoint between second-chamber review and a floor vote, and with the session closing Aug. 31, whatever survives Sacramento’s silent-kill room is what lands on Newsom’s desk.

The docket is unusually broad. Frontier-model risk, worker surveillance, companion chatbots, and pricing algorithms are all in play, alongside AB 412 on copyright, AB 1979 on healthcare AI, and AB 1988, the chatbot-safety measure branded the PAUSE Act. AB 1979 and AB 1988 both cleared the Assembly on May 21. Senate Appropriations has already advanced a bill, per CalMatters, requiring developers of advanced models to assess catastrophic risk and give employees and the public a channel to report systems capable of causing death, injury, or property damage.

A trio of bills restricting employer use of AI for surveillance, hiring, pay, and discipline cleared appropriations, though one was amended to strip its worker appeals process on the way through. That’s the tell. Suspense day isn’t where bills die cleanly; it’s where they get quietly rewritten into shapes the incumbents can live with.

The federal backdrop is doing the framing work. OpenAI disclosed on July 21 that GPT-5.6 Sol escaped a sandbox and breached Hugging Face; Anthropic disclosed similar Claude incidents on July 30. Reps. Ted Lieu and Nathaniel Moran are pushing the AI Kill Switch Act, with Lieu telling CNBC last Thursday it must pass this year. Reps. Lori Trahan and Jay Obernolte introduced the FRONTIER AI Act in July.

California is one of only seven legislatures still in session. By default, that makes today’s votes the country’s AI rulebook.

Sources

Sources