California AI Firing Law Requires Human Review of Dismissals

What happened

California has become one of the first US states to put firm limits on how employers can use artificial intelligence to manage their workforces. Governor Gavin Newsom signed a package of workplace measures that, among other things, prohibits companies from firing employees based entirely on decisions made by AI systems 2. The centerpiece, known as the No Robo Bosses Act, requires that a human review such decisions, with that obligation taking effect in July of next year 1.

The dismissal rule is the headline provision, but the broader package goes further. The new laws also bar employers from using AI to predict workers' emotional states and from collecting neural data, meaning information drawn from the electrical signals of a person's brain or nerves 2. Companies will have to tell workers when layoffs were driven by AI, and AI-based surveillance is banned in workplace bathrooms 2.

Two framings of the same move

Coverage has emphasized different parts of the story. One account centers on the No Robo Bosses Act and its core requirement: no termination decided solely by an algorithm, and a human in the loop starting next July 1. Another treats the firing restriction as one piece of a wider regulatory push. That account presents the laws as a direct response to workers' fears about AI taking their jobs, enabling discrimination, and expanding surveillance on the job 2.

The two accounts don't contradict each other. They differ in scope. The narrower framing makes the policy look like a targeted procedural safeguard. The broader one makes it look like the start of a workplace rulebook for AI. The broader reading seems more accurate. A ban on emotion prediction, limits on neural data, AI-layoff disclosure, and a bathroom surveillance prohibition together amount to more than a single rule about firing.

Why it matters

The significance comes partly from where the laws were passed. California is home to many of the leading companies building AI, and it is moving to regulate the technology's use in the workplace at a time when the federal government has largely stepped back 2. Unions, worker advocates, and some lawmakers pushed for the rules, which mark a shift for a technology that has mostly developed without oversight 2. Some observers expect the laws to serve as a template for broader protections nationwide 2.

The human-review requirement deserves attention because of what it does and does not do. It does not stop employers from using AI in personnel decisions. It stops them from handing the final call entirely to a machine 12. In practice, how much protection this offers will depend on how meaningful that human review turns out to be. A manager who rubber-stamps an algorithm's recommendation satisfies the letter of a "not solely AI" standard more easily than its spirit. How regulators and courts define adequate review will likely decide whether the law has real force.

The disclosure rule on AI-caused layoffs may prove just as consequential in a different way 2. Requiring companies to say when automation drove job cuts would produce public information about how fast AI is actually replacing workers. Right now that question is debated mostly through anecdote and forecast.

The more unusual provisions

The bans on emotion prediction and neural data collection stand out because they address uses of AI that are still emerging rather than widespread 2. Lawmakers who regulate before a practice becomes entrenched avoid the harder task of unwinding it later. The bathroom surveillance prohibition, by contrast, reads as a baseline protection. Its inclusion suggests advocates saw a need to set limits on monitoring that most people would assume already existed 2.

The takeaway

The clearest reading is that California has chosen to regulate AI at the point where it touches people most directly: hiring, firing, and day-to-day monitoring. It has not tried to govern the underlying models. That approach sidesteps some of the fiercest industry opposition to AI regulation while still responding to workers' most concrete fears. Whether the measures hold up will depend on enforcement and on how rigorously "human review" is interpreted once the rule takes effect next July 1. Even so, the state has set out an early marker that other jurisdictions are likely to study, either as a model to copy or as a test case to watch.