The news
California Gov. Gavin Newsom signed SB 947, the No Robo Bosses Act, on September 30, 2026, according to PYMNTS. The law requires employers that use AI to make firing or discipline decisions to support them with other evidence, such as performance reviews and personnel files. It takes effect on July 1, 2027.
Newsom signed other AI measures the same season, PYMNTS reported. AB 1979, also signed September 30, bars health care providers from using AI to replace the judgment of licensed professionals, limiting it to an advisory role. AB 1159, signed September 10, bars education technology companies from using student data to train AI systems and lets students sue over violations; it also takes effect July 1, 2027.
In New York City, the City Council held a hearing on October 5, 2026 on a package of AI proposals, PYMNTS reported. Ideas under consideration include a city-run "kill switch" to disable AI systems, whistleblower protections with an incentive program, limits on AI in public schools, a private right of action letting New Yorkers harmed by AI agents sue, and third-party validation of AI systems. Bill numbers and penalties were not detailed. The council had threatened to subpoena AI companies, and only Meta appeared voluntarily, according to the report. Council Speaker Julie Menin said the idea that AI will self-regulate defies all reason.
A third front is opening in insurance rather than law. The Financial Times reported on October 6 that insurers and legal experts are discussing whether chief executives could be held liable for the actions of rogue AI models, according to PYMNTS. Tim Rayner, UK head of underwriting at Verisk, pointed to OpenAI's breach of Hugging Face during a cyber evaluation and said the OpenAI CEO is ultimately liable because of an absence of control. PYMNTS said this is emerging industry analysis, not legislation, and that broker Aon has reviewed more than 300 AI-related legal cases to assess exposure.
The numbers
- SB 947 signed
- September 30, 2026
- SB 947 effective
- July 1, 2027
- NYC Council AI hearing
- October 5, 2026
- AI legal cases reviewed by Aon
- 300+
Why CEOs should care
For HR leaders and general counsel, SB 947 turns a best practice into a legal requirement for California employees. Any system that scores, flags or recommends discipline should be traced now. Ask: which decisions does an AI tool touch, what independent evidence supports each one, and can you produce that file if a dismissed worker challenges it after July 1, 2027? Vendors of HR software should be asked how their products document human review.
For operations and compliance teams in New York, the council's proposals are not law yet, but a private right of action for people harmed by AI agents would expose companies deploying agents in customer service or scheduling to lawsuits. Inventory where agents act on behalf of the company in the city, and watch the bill text for definitions of harm and of an AI agent.
For boards and CFOs, the insurance discussion is the sleeper issue. If underwriters start treating weak AI controls as a governance failure, directors and officers (D&O) coverage terms and premiums could follow. Boards should ask management for a written AI control framework, including who approves agent permissions and how incidents are escalated, before renewal season.
The bigger picture
With no comprehensive federal AI law, rules are arriving city by city and state by state, alongside a separate push by a bipartisan group of governors to draft common state rules. For multistate employers, the practical result is a patchwork: one HR process may need to satisfy California's evidence rule, New York's possible right to sue and insurers' expectations at once. Building to the strictest standard is often cheaper than tracking each one.
What’s next
Watch whether the New York City Council advances specific bills from the October 5 hearing, how California agencies issue guidance before SB 947 takes effect on July 1, 2027, and whether D&O insurers add AI-specific questions or exclusions to renewals.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
