California has set some of the most detailed rules in the US for how employers may use artificial intelligence on their staff. On September 30, his deadline for acting on this year's bills, Governor Gavin Newsom signed a package of AI laws covering workplace decisions, surveillance, healthcare, legal practice, content provenance and biosecurity.
The centrepiece is a set of worker protections. Employers may not rely solely on AI to discipline or fire employees, and they must be transparent when AI-driven changes lead to mass layoffs. Newsom framed the package around a simple principle: AI should widen opportunity, not cost workers and families.
The Worker Protection Bills
Four bills address AI and surveillance at work:
- SB 947 (McNerney) – governs automated decision systems in employment and bars employers from leaving disciplinary and termination decisions to AI alone
- SB 951 (Gómez Reyes) – requires technological displacement notices, adding transparency when AI systems lead to mass layoffs
- AB 1331 (Elhawary) – sets limits on workplace surveillance, including a ban on monitoring in bathrooms
- AB 1883 (Bryan) – regulates workplace surveillance tools
Together they put a human in the loop for the decisions that matter most to employees, and they bring AI-related job losses into the open instead of hiding them behind generic restructuring announcements.
Beyond the Workplace
The package goes well beyond employment law. Newsom also signed:
- AB 1979 (Bonta) and SB 503 (Weber Pierson) – rules for AI in healthcare services
- SB 574 (Umberg) – stops attorneys, arbitrators and judicial officers from fully handing core legal work to AI
- AB 2392 (Fong) – procurement standards and AI training requirements for public colleges and universities
- AB 2713 (Wicks) and SB 1000 (Becker) – updates to the California AI Transparency Act, including system provenance data for AI-generated content
- SB 1111 (Ashby) – protections around digital replicas
- SB 1159 (Cabaldon) – AI transparency and governance requirements
- AB 1864 (Berman) – oversight of gene synthesis equipment makers and providers, a biosecurity safeguard against AI-assisted misuse
Child safety was also on the list. Newsom signed Adam's Law, a bipartisan measure requiring safeguards for companion chatbots used by children, including suicide and self-harm crisis protocols, parental controls and alerts when safety settings are switched off.
These signings follow a busy month. In early September Newsom signed SB 1050, which requires disclosure when ads use AI-generated synthetic performers. California has also been building an independent oversight regime for frontier AI companies, including work toward an AI kill switch under an earlier executive order. Governors in Oregon and Illinois have since issued similar AI executive orders.
A National Patchwork Grows
California is not acting alone. On October 1, Connecticut's SB 5, the C.A.R.T. Act, takes effect. It requires clear consent before AI subscription services auto-renew and protects whistleblowers at AI companies who report risks to public health or safety. One legislative tracker counts 84 new AI laws enacted across 27 states so far in 2026.
At the federal level, regulation is mostly voluntary. This week the White House announced a "morally binding" safety accord with six frontier companies, with no enforcement mechanism. In Congress, Reps. Jay Obernolte and Lori Trahan have released a discussion draft of the bipartisan Great American AI Act, which would make third-party audits of frontier AI mandatory. Until something like that passes, states such as California are writing the rules that companies actually have to follow.
Why It Matters
California's labour market and its concentration of AI companies give its laws influence well beyond the state. Employers operating nationally often adopt the strictest standard rather than maintain separate HR processes for each state, so human review of AI-driven firings and layoff disclosure may become standard practice across the US.
For HR and legal teams, the practical steps are clear:
- Audit automated decision systems used in hiring, performance management and termination, and document where humans review outcomes
- Map workplace monitoring tools against the new surveillance limits
- Prepare disclosure processes in case AI adoption leads to workforce reductions
- Check AI content pipelines for provenance-data and digital-replica compliance
The open question is enforcement. Previous California AI laws have drawn lawsuits and pushback from industry, and the federal government has repeatedly argued for pre-empting state AI rules. For now, though, California has turned the abstract debate about AI and jobs into specific obligations that employers can be held to.
