California Becomes First State to Require a Human Sign-Off Before AI Can Fire You
Gov. Gavin Newsom signed SB 947, the 'No Robo Bosses Act,' into law on September 30, making California the first state to bar employers from relying solely on an automated system to fire or discipline a worker. The law, authored by Democratic state Sen. Jerry McNerney, requires that any AI-flagged termination or discipline decision be independently corroborated by a human using personnel files, manager evaluations or peer reviews before it takes effect, and that affected workers get written notice describing what data the system used against them, plus a human contact who can explain the decision.
The signing closes a year-long fight that says as much about the limits of California's appetite for AI employment regulation as it does about its ambitions. Newsom vetoed an earlier version of the bill, then numbered SB 7, in October 2025, citing concerns that its advance-notification requirements were too broad. McNerney reintroduced a revised version in February 2026 that dropped the advance-notice mandate and, according to reporting from Techstrong.ai, eliminated protections that would have covered gig workers' 'deactivation' decisions — the mechanism platforms use to cut off app-based drivers and couriers. The version that passed also carves out exceptions for national-security and defense-related uses.
A narrower win, but a real one
"No worker should ever be fired or disciplined by a machine, AI or not. AI must remain a tool controlled by humans." — Jerry McNerney, California State Senator
The law's enforcement teeth are modest: violations carry a $500 civil penalty enforced by the state Labor Commissioner, with no private right of action, according to legal analysis from Mintz covering the bill's development. That's a far cry from the uncapped litigation exposure business groups feared from the original draft. The California Federation of Labor Unions, AFL-CIO, which sponsored the bill, still counted it as a landmark; the Chamber of Progress, a tech industry group, opposed the final text anyway, arguing that ambiguity in what counts as 'primarily relying' on an automated system could chill adoption of workplace AI tools altogether, per Techstrong.ai's reporting.
SB 947 takes effect July 1, 2027 — giving employers roughly nine months from signing to build the human-review layer the law requires, longer than the typical January 1 start date for most California statutes. That runway matters: the law's core obligation isn't a disclosure requirement employers can satisfy with a privacy-policy update, it's a process requirement that forces a human into the loop on every covered personnel action, which means new workflows, recordkeeping and likely new headcount in HR and legal compliance functions at scale.
The law also stakes out a different model than the other state furthest along on workplace AI, Illinois, whose amended Human Rights Act (effective January 1, 2026) requires employers to notify workers when AI is used in employment decisions and reaffirms that such use can't produce discriminatory outcomes — but stops short of requiring a human to independently verify any AI-driven firing or discipline decision. California's law is the first to convert 'human oversight' from a disclosure obligation into a verification mandate, and other legislatures weighing their own versions next session will be watching whether that distinction holds up once it's actually enforced.