Newsom signs 'No Robo Bosses Act': humans must sign off on AI-driven firings
Governor Gavin Newsom signed the "No Robo Bosses Act" (SB 947) on September 30, 2026 — according to the Times of India, the first law in a US state to ban terminations and disciplinary cases based solely on AI. The law, which requires human review and notification of affected workers, takes effect July 1, 2027.
What the law does
SB 947 has, according to IT Pro, three components. First, employers are barred from relying solely on an automated decision system when terminating or disciplining workers. Second, human oversight and verification are required when such systems are used to assist in these decisions. Third, the affected worker must be informed if a company uses automated systems in termination or disciplinary cases.
The review requirement has concrete content: the human verifier must, according to Firstpost, assess more than just the system's recommendation — including managerial assessments, personnel files, and peer evaluations. The transparency obligation means workers are told that an automated system was used, what employee data the system considered, and who the human contact who can explain the decision is.
An ambiguous threshold
The sources describe the review trigger differently. Firstpost writes that human review is required when AI plays a "significant role" in the decision, while Times of India, citing CNBC's coverage, renders the law's standard as "primarily" — that is, employers relying primarily on AI results must have a human verifier confirm the decision. The difference matters practically: a "significant role" standard captures more employer use than a "primarily" standard. The exact statutory threshold should be verified against the bill text, which is not reproduced in any of the available sources.
From veto to enactment
The law is a narrower follow-up to SB 7, which Newsom vetoed in 2025. In his veto message, he said, according to IT Pro, that the effort should target misuse, not impose "unfocused notification requirements on any business that uses even the most harmless tools." The law was sponsored by the California Federation of Labor Unions, AFL-CIO, and Senator Jerry McNerney first introduced it in 2025 after sustained pressure from organized labor, writes Times of India.
The narrower version won broad political support: it passed the Assembly 53–14 and the Senate 28–10, according to the California Employment Law Report (Anthony Zaller, September 4, 2026). The July 1, 2027 effective date is now confirmed by the signing.
What was cut from SB 7?
Here the sources disagree. IT Pro writes that a notification requirement — to notify employees when an AI system would affect their work — was removed, along with protections for gig workers. AGIRight, summarizing CalMatters' coverage of the governor's press release, states instead that the signed version dropped an appeals process, workers' right to sue, and coverage of contractors. The two descriptions may partly reflect the same narrowing from different angles, but they do not fully overlap. The exact scope of the difference between SB 7 and SB 947 cannot be fully determined without the bill text, and readers should treat the various lists of cuts as partially confirmed.
Same day: a package of AI laws
Newsom signed SB 947 on September 30, 2026 — the final day of his signing window — along with several other AI-related labor laws, according to AGIRight, citing the governor's press release:
- SB 951: Employers must disclose when a mass layoff, relocation, or termination is due to an AI system.
- AB 1331: Ban on workplace surveillance tools in bathrooms.
- AB 1883: Ban on using AI to infer workers' emotional states or collect neural data.
Why the law matters
Newsom justified the signing by pointing to the fallibility of AI systems: "No worker should ever be fired or disciplined by a machine, AI or not," he said, according to Times of India (reproduced via CNBC). "Artificial intelligence has the potential to increase productivity, but it has also made mistakes and misjudgments and shown bias."
McNerney, the law's author, framed it as a question of control: "AI must remain a tool controlled by humans, not the other way around," he said, according to Firstpost.
This is the first law of its kind in any US state, according to Times of India's coverage, and it comes as employers increasingly adopt automated decision systems in personnel processes.
Enforcement and the unresolved
The law is enforced by California's labor commissioner, the state attorney general, or local prosecutors, according to IT Pro. Similar laws are being drafted in a handful of other states, making California's implementation a possible template.
Several questions remain open before the effective date:
- The threshold for human review: Whether the law uses a "significant role" or "primarily" standard, and how it is to be interpreted in practice, is not settled in the available source material.
- Penalties: What fines or other consequences violations may carry has not been reported.
- The scope of the cuts: As described above, the sources differ on exactly what was removed from SB 7.
- Practical compliance: How employers are to document that the human verifier actually used the additional information remains to be seen.
Organized labor received the legislative package favorably but called it incomplete, according to CalMatters' coverage reproduced by AGIRight — a sign that the questions of contractors, right to sue, and further protections will likely return in future bills.
Heading toward July 1, 2027, the decisive factors will be how employers adapt their termination and disciplinary processes, and how enforcement authorities choose to interpret the law's core requirement of human verification.