California's 3 New AI Employment Laws Reshape Litigation Risk

Employment counsel now have three new California statutes to fold into AI governance and workplace policy reviews before year-end. The measures ban using AI to infer employee emotional states from biometric data, forbid AI as the sole basis for termination, and mandate written notices when AI drives mass layoffs, creating private-litigation and state enforcement exposure for employers and HR-tech vendors.

On Wednesday, September 30, 2026, California Governor Gavin Newsom signed three employment-focused artificial intelligence bills that establish some of the nation's most concrete workplace guardrails against AI-driven decisions and surveillance. The measures ban employers from using technology to predict a worker's emotional state through biometric data, require written notice when artificial intelligence is responsible for mass layoffs, and prohibit employers from relying solely on AI to fire an employee. Simultaneously, Newsom signed an executive order directing state agencies to keep using the phrase 'artificial intelligence' rather than President Donald Trump's preferred 'super intelligence'. The signing came on Newsom's final day to approve or veto legislation before he terms out in January, amplifying both the immediate compliance stakes and the political signal sent to Washington.

“

On Wednesday, September 30, 2026, California Governor Gavin Newsom signed three employment-focused artificial intelligence bills that establish some of the nation's most concrete workplace guardrails against AI-driven decisions and surveillance.

California's decision has immediate consequences for employers and technology vendors operating in the state. The biometric-emotion ban targets a rapidly growing category of workplace monitoring tools marketed for engagement, safety, and productivity, but often built on facial expression analysis, voice analysis, or physiological sensors. By cutting off the use of biometric data to infer emotional states, the law pushes California beyond narrower data-privacy statutes toward substance-based limits on AI inference. Companies that currently deploy affective computing in hiring, call centers, warehouse safety, or performance management will need to audit vendors, disable noncompliant features, and document human oversight. The layoff-notice requirement adds a new disclosure duty layered on top of California's existing mass-layoff notification regime: if AI is responsible for selecting roles or individuals in a reduction in force, affected workers must receive a written notice making that role explicit. This creates both operational complexity and a future evidentiary record in unemployment hearings and discrimination litigation. The prohibition on AI-only termination is perhaps the most consequential for legal risk, requiring a human decision-maker with meaningful authority to be involved in firing decisions. Employers will need written policies, training, and record-keeping to prove that any adverse action was reviewed by a qualified human.

Newsom framed the package as a direct response to federal inaction. At an hourlong press conference Wednesday, he sharply criticized Trump for failing to enact comprehensive federal AI regulation while industry leaders warn about risks. Trump responded Tuesday by announcing that top technology executives signed a voluntary accord to 'self-police' AI development. The governor also pointed to earlier California actions: a law signed this month requiring operators of AI chatbots to perform risk assessments before rollout and an executive order to consult with experts on further oversight. The state-federal fight is now not just philosophical but operational, because the executive order on terminology instructs California agencies to reject the administration's renaming effort. The conflict sets up possible intergovernmental disputes over federal contracts, research programs, and educational materials that use the contested terminology.