A year ago, California tried to ban the "robo boss" and Governor Gavin Newsom vetoed it. On September 30, 2026, he signed a narrower version of the same idea into law — and in doing so, created what its author calls the nation's first statute requiring human oversight whenever an automated system is used to discipline or fire an employee. Here's what actually changed between the version he rejected and the one he just signed, and what it means if you work for, or run, a California business.
Quick summary: Newsom vetoed the original "No Robo Bosses Act" (SB 7) on October 13, 2025 for being too broad. On September 30, 2026, he signed its replacement, SB 947, which bars employers from relying solely on an automated decision system to discipline or terminate workers, requires human corroboration when such a system is the primary basis for the decision, and mandates written notice to the affected employee. A companion bill, SB 951, expands California's WARN Act to require notice when AI or automation substantially causes a mass layoff. SB 947 takes effect July 1, 2027.
Why the first version got vetoed
The original No Robo Bosses Act, SB 7, introduced by state senator Jerry McNerney, would have required employers to notify workers before using any automated decision system that could affect their employment — a definition broad enough, Newsom said in his October 2025 veto message, to sweep in "even the most innocuous tools" alongside genuinely consequential ones. He said he shared the underlying concern but called the bill's restrictions "overly broad." That veto could have killed the idea for the session; instead, the bill's backers narrowed its scope and brought it back (Source: Mintz law firm analysis).
What SB 947 actually requires
The rewritten bill, SB 947, drops the broad advance-notice requirement for every automated tool and instead focuses specifically on consequential disciplinary and termination decisions. Employers can't rely solely on an automated decision system (ADS) to discipline or fire someone, and can't use one to infer an employee's protected characteristics or to predict or retaliate against someone for exercising a legal right. Where an employer primarily relies on an ADS for such a decision, a human reviewer must examine and corroborate it using independent information — things like a manager's own evaluation, peer feedback, or the employee's personnel file — and if the system's output turns out to be inaccurate or incomplete, it can't be the basis for discipline or termination. Affected employees get a written notice that the system was primarily used, a meaningful description of what employee data the system relied on, and a named human point of contact who can explain the decision (Source: SHRM).
SB 951: when AI costs jobs, workers get a notice
Signed the same day, SB 951 extends California's existing WARN Act — which already requires advance notice of mass layoffs and plant closures — to specifically cover job losses caused in whole or substantial part by an AI system or other automated technology that replaces or automates employment positions. In practice, that means a company can't quietly attribute a mass layoff to "restructuring" if automation is the real driver; it has to say so in the legally required notice (Source: Governor's office announcement).
- SB 947 takes effect July 1, 2027 — employers have roughly nine months from signing to build human-review processes into any AI-assisted discipline or termination workflow.
- The human reviewer must use genuinely independent information, not just rubber-stamp the AI system's own output.
- Employees gain a specific, enforceable right to ask what data the system used about them.
- SB 951 means a mass layoff notice now has to disclose if AI or automation substantially caused it, not bury that behind generic language.
What this means if you don't work in California
California employment law has a long history of becoming a template other states borrow from once it survives its first legislative session, and AI-in-the-workplace rules look likely to follow that pattern — several other states have similar bills in committee. For workers anywhere, the practical takeaway is simpler: if your employer ever cites "the system flagged you" as the reason for discipline, California's law (and the direction other states are heading) says that's not supposed to be allowed to stand on its own, and you're entitled to ask what a human actually reviewed.
If you manage people in California and use any AI tool for performance tracking, scheduling, or discipline recommendations, SB 947's July 2027 effective date isn't far off — it's worth auditing now whether a human actually reviews independent information before any AI-flagged termination or discipline, rather than waiting until the law takes effect to find out you don't have a compliant process.
01When does California's AI employment law take effect?
SB 947, the No Robo Bosses Act, takes effect July 1, 2027. It was signed by Governor Newsom on September 30, 2026.
02What's different between SB 7 (vetoed) and SB 947 (signed)?
SB 7 would have required advance notice before using almost any automated workplace technology; Newsom vetoed it in October 2025 as overly broad. SB 947 narrows the focus specifically to automated systems used in consequential discipline and termination decisions, requiring human corroboration and written notice rather than blanket advance disclosure.
03Can an employer still use AI to help with firing decisions in California?
Yes, but not as the sole basis. If an employer primarily relies on an automated system for a discipline or termination decision, a human reviewer must corroborate it using independent information, and the employee must receive written notice plus a human contact who can explain the decision.
04What is SB 951?
A companion law, signed the same day as SB 947, that expands California's WARN Act to require employers to disclose when a mass layoff, relocation, or termination is caused in whole or substantial part by an AI system or other automating technology.
Whether you're trying to understand a new workplace AI policy, draft a compliant notice, or just want a second opinion on something your employer told you, LumiChats gives you access to Claude, GPT, Gemini and 40+ other models to cross-check how each one reads a law like this — useful when the stakes are as personal as your own job.
