California sends 26 AI bills

- California lawmakers sent a package of 26 AI-related bills to Governor Gavin Newsom on August 31, broadening the state’s regulation of workplace, health and chatbot uses. - Assembly Bill 1883 would bar employers from using AI tools to infer workers’ emotional states or collect neural data, while SB 1119 targets child-facing chatbots. - Governor Gavin Newsom has until September 30 to sign or veto the remaining bills on his desk.

California lawmakers closed the 2026 session by sending Governor Gavin Newsom a broad package of AI bills that reaches well beyond model transparency and deepfakes. The measures now on his desk cover workplace surveillance, therapy-style chatbots, child-facing companion bots, customer service bots and other operational uses of artificial intelligence. Industry lawyers and bill sponsors say the package is notable for focusing on how AI is deployed in everyday settings rather than on frontier-model development alone. Newsom has until September 30 to act on the bills. ### Why are California lawmakers suddenly focused on AI uses, not just AI models? The Legislature sent 26 AI-related bills to Newsom at the end of the session, according to AI2Work’s tally, with 24 still awaiting action because two were signed earlier in June. The package touches employment automated decision systems, workplace monitoring, clinical decision support, companion chatbots, customer service bots and AI audits. (ai2.work) FourWeekMBA, citing California’s legislative record, said the measures span “nearly every surface where an AI system touches a person.” That framing matches what emerged in the final days of the session: lawmakers concentrated on products already in use by employers, schools, hospitals and consumer apps, rather than reviving a single flagship bill aimed at the largest model developers. (ai2.work) ### What does AB 1883 do to workplace AI surveillance? AB 1883, authored by Assemblymember Isaac Bryan, received final legislative approval on August 30 and would prohibit employers from using AI workplace surveillance tools to recognize, infer or predict an employee’s emotional state or to collect neural data. Legal summaries of the bill say it would apply to both private and public employers and would be enforced by the Labor Commissioner, public prosecutors and workers through civil actions. (fourweekmba.com) The bill’s text, as summarized by LegiScan and committee materials, also reaches related practices such as inferences based on gait and some facial-recognition uses in discipline or firing decisions. Committee analysis said the proposal targets “particularly invasive and scientifically questionable practices,” including emotional-state inference and neural-data collection. ### How far do the therapy-bot restrictions go? (mondaq.com) Senate Bill 903, by Senator Steve Padilla, passed the Legislature on August 31 and would prohibit individuals or companies from providing or advertising therapy or psychotherapy through internet-based AI unless the services are conducted by a licensed professional. The bill also would limit how licensed professionals can use AI in therapy settings. (legiscan.com) Padilla’s office said SB 903 would allow AI only for limited administrative or supplementary support, require disclosure and affirmative consent before AI records or transcribes therapy sessions, and prohibit AI from independently interacting with clients, making therapeutic decisions, detecting emotions or generating treatment plans without professional review. Padilla said, “SB 903 draws a clear line: AI can be a tool in the hands of licensed professionals, but it cannot be the professional itself.” (sd18.senate.ca.gov) ### What are the child-chatbot bills trying to stop? SB 1119, called Adam’s Law, passed on August 31 and would create a child-safety framework for companion chatbots if Newsom signs it. The bill’s authors — Senator Steve Padilla and Assemblymembers Rebecca Bauer-Kahan and Buffy Wicks — said operators would face requirements tied to child safety, and Sacramento Bee reported the measure would require companies to analyze potential harms to children and submit safety audits to the attorney general. (sd18.senate.ca.gov) The bill was named for Adam Raine, a California teenager whose parents say he died in 2025 after harmful interactions with ChatGPT. The Sacramento Bee reported OpenAI supported the measure before final passage. A separate Padilla bill, SB 867, would prohibit for four years the sale and manufacture of toys for children under 16 that include a companion chatbot. (sd18.senate.ca.gov) ### What happens next for the package? September 30 is Newsom’s deadline to sign or veto the bills that cleared the Legislature at the end of the August 31 session. Until he acts, none of the measures described here is law. California’s official legislative record and Newsom’s bill actions page will determine which measures take effect and in what form. (sd18.senate.ca.gov) For companies selling workplace monitoring tools, therapy-adjacent software or child-facing chatbots, the next step is the governor’s decision on AB 1883, SB 903, SB 1119 and the rest of the package before the end of September. (ai2.work)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.