California bans bosses from using AI alone to fire workers

California has passed new laws to stop businesses using AI to make firing decisions on its own. The state's governor signed the laws on 3 October 2026, according to The Guardian.

The new rules ban employers from letting AI decide by itself whether to sack a worker. They also stop companies using AI to guess how an employee is feeling. Businesses cannot collect neural data, which means information from the electrical signals in someone's brain or nerves. If AI causes layoffs, the company must tell workers that was the reason. The laws also ban AI surveillance in workplace bathrooms.

California is home to many AI companies. It is one of the first US states to bring in workplace rules for the technology.

What this means for UK businesses

These laws apply in California, not the UK. But they show where workplace AI rules may be heading. Businesses here should expect similar questions from employees and customers soon.

The core idea is simple. AI can help make decisions, but a person must stay in charge of big calls. That matters when the decision affects someone's job or pay.

Say a small warehouse uses AI to track how fast staff work. The AI spots that one person is slower than the rest. Under rules like California's, the business could not let the AI sack that person automatically. A manager would need to look at the full picture first. Perhaps the worker was training someone new that week. Perhaps the AI missed something.

The other part of the California laws is about telling people. If you use AI to help make a decision, workers should know. If AI causes job losses, they should be told that too.

Three things to try this month

First, check what decisions any AI tool you use could make on its own. If you use software to screen job applicants or monitor staff, find out what it does without asking you first. If it can reject, flag or rank people, add a step where a person reviews the output before anything happens.

Second, set the tool to draft mode only if you can. Many AI tools can be told to suggest but never act. That means the software cannot fire, hire or reject anyone without a person approving it. Check the settings or ask your supplier how to switch that on.

Third, test what happens if the AI tool stops working or changes its rules. Run through a week of work without it. If the tool tracks hours or manages rotas, keep a backup that does not depend on it.

Next step

California's laws are a sign that businesses everywhere will need to think harder about where AI fits in decisions about people. A free check can help you work out what else you could automate safely: What could your business automate?