Menu

California's No Robo Bosses Act is law: AI alone cannot fire or discipline workers

Comingwave team · 7 minute read · published

A manager and an employee sit across a table reviewing a paper file, with a blurred chart on a nearby screen.

California's No Robo Bosses Act is now law. On 30 September 2026 Governor Gavin Newsom signed Senate Bill 947, which its author, Senator Jerry McNerney, calls the No Robo Bosses Act of 2026, as part of a package of AI bills. From 1 July 2027 an employer covered by the law may not rely solely on an automated decision system to discipline or dismiss a worker. Where such a system is the main basis for the decision, a person must check it against other evidence and the worker must be told in writing. It sits in California's Labor Code, and its definition of employer includes labour contractors. This article is general information, not legal advice.

What happened

The Governor's office announced the signing in a release titled California's nation-leading AI framework just got stronger. It lists 13 bills. Four are workplace measures: SB 947 on automated decision systems, SB 951 on notice of technological displacement, and AB 1331 and AB 1883 on workplace surveillance. The release quotes the Governor: "AI should expand opportunity – not come at the expense of workers and families."

The chaptered text of SB 947 on the California Legislature's website shows it was approved by the Governor and filed with the Secretary of State on 30 September 2026 as Chapter 859. It adds a new part to the state's Labor Code on automated decision systems in the workplace.

Senator McNerney's office, in its release on the signing, calls it the first such law in the nation and says the bill was sponsored by the California Federation of Labor Unions. The senator's summary: "AI must remain a tool controlled by humans, not the other way around."

Key details

QuestionWhat SB 947 says
When does it start?The new part becomes operative on 1 July 2027
What is an automated decision system (ADS)?A computational process derived from machine learning, statistical modelling, data analytics or AI that issues simplified output, such as a score, classification or recommendation, used to assist or replace human discretionary decision-making and that materially impacts people. Spam filters, firewalls, antivirus software, calculators and databases are excluded
Who is an employer?A person who directly or indirectly, or through an agent, employs or exercises control over the wages, hours or working conditions of any person. Labour contractors and government bodies are included
What can an ADS never be used for?To prevent compliance with or violate labour, occupational health and safety, employment or civil rights laws; to infer an employee's protected status; or to predict and take adverse action against a worker for exercising legal rights
Discipline and terminationAn employer "shall not rely solely on an ADS". If it primarily relies on ADS output, it must direct a human to corroborate the decision
PenaltyA civil penalty of $500 per violation, enforced by the Labor Commissioner or a public prosecutor

Three duties follow when an ADS is the primary basis for a disciplinary or termination decision:

  1. Corroboration by a person. The reviewer uses the data behind the output or other supporting information. The law gives examples: supervisory or managerial evaluations, personnel files, the employee's work product, peer reviews and witness interviews. If the output cannot be corroborated, or the reviewer finds it inaccurate, incomplete or misleading, the employer must not use it for the decision.
  2. A written notice. At the time the employee is told of the decision, the employer must give a separate, plain-language notice stating that it primarily relied on an ADS, that a human reviewed and corroborated the output, who to contact for more information, and that retaliation is prohibited.
  3. Access to a description of the data. The employee may request, and the employer must provide, a meaningful and objective description of the employee's own data used by the system, with other people's personal information anonymised.

There are limits. The law does not apply where a collective bargaining agreement explicitly waives it and provides protection from algorithmic management, or where an ADS is required for certain federal aerospace, defence and national security work. An employer subject to California's consumer privacy law must still follow the privacy regulator's rules on automated decision-making technology.

Why it matters

Three features make this more than a local employment rule.

The definition is broad. It is not limited to products sold as AI. A productivity score from a monitoring tool, a risk flag from a fraud model or a ranking from a rostering system can all be "simplified output" that assists a decision about a person. The definition of employer has no minimum headcount.

The burden of proof shifts. In an enforcement action, once it is shown that an ADS was used in the decision, the employer must demonstrate that it did not primarily rely on the system or that it complied with the corroboration and notice rules. Good records are the practical answer to that.

It is narrower than the name suggests. The corroboration and notice duties attach to discipline and termination. They do not regulate hiring, pay or promotion decisions, although the three outright prohibitions apply to any use of an ADS by an employer.

What this means for businesses

An Australian business with employees in California, directly or through a labour contractor, should ask a local employment adviser whether and how the law applies to it. Software companies meet it indirectly: if you sell workforce, HR or performance tools to employers there, expect customers to ask whether the product supports review, records and notices.

  • List the systems that score or flag staff. Time and attendance, productivity monitoring, call quality scoring, delivery or driver ratings, fraud and safety alerts.
  • Ask of each one: could a warning, suspension or dismissal rest mainly on its output?
  • Name the reviewer. Decide who checks the output, what other evidence they must look at, and where they record it.
  • Prepare the notice. A stand-alone, plain-language template with a contact person, ready before 1 July 2027.
  • Be able to describe the data. Check that each system can export, in readable form, the data it used about one employee without exposing customers or colleagues.
  • Put it in supplier contracts. Ask vendors to confirm their product can provide the underlying data and an explanation of the score.
  • Brief managers. The law prohibits retaliating against an employee for using their rights under it, such as asking for the description of their data.

The same questions come up earlier in the employment relationship, at hiring. That is the stage where a recruitment service such as Talynx works with employers, and it is reasonable for clients to ask how any automated assessment of candidates is reviewed by a person.

On the systems side, the work is mostly plumbing: connecting HR, rostering and monitoring tools so a reviewer can see the evidence in one place, and keeping an audit trail of who decided what. That is the kind of job we do under business systems and integrations, with data dashboards for the people who review and custom software where a workflow needs a review step built in. If you would like a second opinion on how your workforce systems fit together, send us an enquiry.

Key takeaways

  • Governor Newsom signed SB 947, the No Robo Bosses Act of 2026, on 30 September 2026; it becomes operative on 1 July 2027.
  • Employers may not rely solely on an automated decision system to discipline or dismiss a worker.
  • Where a system is the primary basis, a human must corroborate it, and the worker gets a written notice and a right to a description of their data.
  • The penalty is $500 per violation, and the employer carries the burden of showing it complied.
  • The definition of an automated decision system covers scores and flags from everyday workforce tools, not only products marketed as AI.

Frequently asked questions

When does the No Robo Bosses Act take effect?

The new part of the Labor Code added by SB 947 becomes operative on 1 July 2027.

Does SB 947 ban AI in workplace decisions?

No. It bans relying solely on an automated decision system for discipline or termination, and sets conditions where a system is the primary basis.

Does the law cover hiring decisions?

The corroboration and notice rules apply to disciplinary and termination decisions, not to hiring. The three outright prohibitions, such as using a system to infer an employee's protected status, are not limited to those decisions.

What is the penalty for breaching SB 947?

An employer that violates the new part is subject to a civil penalty of $500 per violation. The Labor Commissioner or a public prosecutor can enforce it.

Does it apply to Australian businesses?

The law is part of California's Labor Code. Its definition of employer covers a person who employs people directly, indirectly or through an agent, and includes labour contractors. An Australian business with employees in California should ask a local employment adviser whether and how it applies. For a business with no workforce there, it is a picture of the questions workplace software is starting to face.

Need help with your business technology?

Tell us what you need. We reply within one business day.

Get a free quote