On Thursday Governor Gavin Newsom signed a package of statutes that bar employers from making termination decisions solely with artificial intelligence, from using AI to infer workers’ emotional conditions, and from gathering neural data such as brain-wave signals. The measures also compel firms to tell employees when a layoff is driven by AI and forbid AI-based monitoring in restroom areas. The action arrives as the federal government has taken a largely hands-off stance toward AI regulation.
The legislation responds to growing employee anxiety that AI could replace jobs, amplify bias, and increase surveillance on the job floor. Unions, labor advocates and a number of legislators pressed for the rules, marking a regulatory shift for a technology that has largely evolved without oversight. California, home to many leading AI developers, thus becomes one of the first states to impose a comprehensive set of workplace protections targeting artificial intelligence.
Several other states have already enacted narrower AI-related workplace statutes, including Colorado, Connecticut, Illinois and Texas. Those measures address specific uses of the technology but do not match the breadth of California’s package. According to labor advocate Gonzalez, additional bills are now being prepared for consideration across the country, indicating a growing legislative appetite for AI governance in employment contexts.
The new rules target surveillance tactics such as heat-map tracking of employee movements and monitoring of bathroom break duration, practices that have drawn complaints from workers at Amazon warehouses. They also curb systems that assess emotional cues, exemplified by a program at Kaiser Permanente that graded nurses’ tone of voice during patient interactions. By outlawing these specific applications, the statutes aim to limit invasive data collection in the workplace.
Union groups such as the California Federation have begun monitoring AI product launches, using market availability as an indicator of workplace deployment, Gonzalez explained. The federation also intends to revive a previously failed bill that would require employers to disclose any AI tools used on the job. “We don’t know all the places companies are using AI, and that is and should be scary,” Gonzalez said, underscoring the need for transparency.
Legal scholars view the legislation as a notable advance but note its limited scope. Robin Feldman, director of the AI Law & Innovation Institute at UC San Francisco, called the measures a “key step” for worker advocates while acknowledging implementation challenges. Employment-law partner Danielle Ochs warned that requiring ten compliance checkpoints per AI tool may not reflect real-world practices, suggesting broader “guardrails” would better align with how firms actually use the technology.
The statutes appear amid record AI spending by major tech firms and significant workforce reductions. In June, Meta halted a program that logged employee computer activity to train its models, and a month later a group of workers sued the company alleging AI tools targeted staff on disability accommodations or medical leave for termination. Simultaneously, OpenAI and Anthropic have urged legislators to slow AI development, reflecting broader safety concerns that intersect with labor issues.
Advocates say the California package marks only the beginning of a broader effort to mitigate AI’s impact on employment. Gonzalez emphasized that the laws are a first step toward addressing potential harms, while acknowledging that their ultimate effectiveness will depend on enforcement and future legislative refinements. The state’s approach may set a template for other jurisdictions seeking to balance technological innovation with worker protections.