California employers will soon need a person to stand behind any artificial intelligence-driven decision to fire or discipline a worker. Gov. Gavin Newsom signed SB 947, the No Robo Bosses Act, on Sept. 30 as one of 13 bills that day, according to the governor’s signing announcement. The law bars employers from relying only on AI when they discipline or fire someone.
SB 947 is one of three new California laws that limit how much authority companies can hand to artificial intelligence. AB 1979, signed the same day, keeps licensed clinicians in charge when healthcare providers use AI. AB 1159 bars education technology companies from training AI on student data. Newsom signed that law Sept. 10 in a separate package of 13 bills on children and technology, according to the governor’s Sept. 10 announcement.
Under SB 947, a manager who relies mainly on AI to discipline or fire someone must back up the decision with other evidence, such as performance reviews and personnel files, HR Dive reported. The law takes effect July 1, 2027.
Newsom vetoed an earlier version of SB 947 in October 2025, saying it was too broadly written, PYMNTS reported. The new measures put “guardrails on AI at work to ensure human oversight,” Lorena Gonzalez, president of the California Federation of Labor Unions, AFL-CIO, said in the governor’s announcement.
Clinicians Keep the Final Call on Healthcare AI
AB 1979, titled Health Care Services: Artificial Intelligence, bars hospitals and other healthcare providers from using AI to replace a licensed professional’s judgment, the California Nurses Association, which sponsored the bill, said in a release. Technology “should not replace human decision-making in health care,” Assemblymember Mia Bonta, who wrote the bill, said in the same release. The law limits AI to an advisory role, KQED reported.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
Some hospitals already work that way. An AI scribe has drafted notes for more than 900 patient visits at University Hospital at Downstate in Brooklyn, and physicians approve each one before it enters the record, PYMNTS reported. A companion bill, SB 503 by Sen. Akilah Weber Pierson, targets the companies that build these tools. It requires developers of AI for diagnosis and treatment to take reasonable steps to reduce bias, the governor’s office said.
Student Records Leave the AI Training Pipeline
AB 1159 bars education technology (EdTech)companies from using student data to train generative AI or build other artificial intelligence systems, according to the bill. The ban covers personal information the companies collect through their websites and apps, including identifiers that track a student over time.
The law also creates the Higher Education Student Information Protection Act, which takes effect July 1, 2027. About 2.9 million California college students had no protections like those that already covered younger students, Privacy Rights Clearinghouse, the nonprofit that sponsored the bill, said in a fact sheet. The group calls the measure the California Learner Personal Information Protection Act.
AB 1159 limits what companies can collect, too. EdTech firms can’t gather or use sensitive information such as a student’s immigration status, the fact sheet said.
Students have a new way to enforce the rules. A student harmed by a company’s violation can sue that company, according to the bill.
California first regulated how EdTech firms use student data in 2014, CalMatters reported. That law banned selling student data and using it for targeted ads. It did not address AI training.
The White House has taken a different approach. President Donald Trump said Sept. 29 that executives from OpenAI, Anthropic, Google and Meta had signed a voluntary accord to self-police AI development, leaving oversight to the model makers themselves, PYMNTS reported.
Add as Preferred Source
