California limits AI firing, mood tracking and hidden AI layoffs

California Governor Gavin Newsom has signed three workplace laws aimed at AI. Software alone will no longer be allowed to decide that a worker gets disciplined or fired. Employers also face limits on AI that tries to read workers' emotions. And big layoffs caused by automation will have to be labeled as such.

What changed

SB 947 takes effect on July 1, 2027. It bars employers from relying only on an automated decision system, meaning software that makes or drives decisions, when they discipline or fire someone. If the software is the main basis, a person must check the underlying evidence. That can include peer reviews, witness interviews and personnel files. The aim is real review, not a manager clicking approve.

Workers must get written notice when such a system was primarily used. They can also ask for a description of the data it relied on.

Surveillance limits

AB 1883 targets AI tools that watch employees. It restricts using them to analyze how workers feel or to predict their emotions. It also restricts collecting data about workers' brain or nervous-system activity.

Employers that break the law can face enforcement action and civil penalties of $500 for each violation.

Layoffs and AI

SB 951 changes California's WARN notice rules. A WARN notice is the advance warning employers must give before certain mass layoffs, closures or relocations. Starting January 1, 2028, if those cuts happen in whole or in substantial part because of AI or other automation, the notice must include extra details:

• How many layoffs are substantially attributable to automation
• The occupations and locations affected
• The job functions being automated
• The category or type of technology involved

The notice must also call the event a "technology displacement." The law does not ban replacing workers with AI. It makes the role of technology visible.

How other places compare

California joins Colorado, Illinois, Maine and New York City in setting rules for automated employment decisions or electronic monitoring, though each takes a different path. Colorado, starting January 1, 2027, will require notices about AI use in employment decisions. Workers hit by a negative decision can request their personal data and ask for human review.

Illinois focuses on hiring. Employers using AI to analyze video interviews must explain how it works, say what it evaluates and get consent. Applicants can ask for their videos to be deleted. New York City requires bias audits within the past year for automated hiring tools, public audit details and notice to candidates.

What is not clear yet

The material does not give a start date for AB 1883. Newsweek said it has asked the California Federation of Labor Unions for comment. Employers with California staff should expect to review how they use automated tools in firing, monitoring and layoffs.