Estimated reading time: 5 minutes · Last updated:
As first reported by The Guardian, Governor Gavin Newsom signed a suite of California laws that restrict how employers can use artificial intelligence in the workplace. The statutes ban relying solely on AI to fire workers, bar using AI to predict emotional states, forbid collecting neural data and prohibit AI surveillance in workplace bathrooms. They also require companies to notify employees when layoffs were caused by AI. The measures target surveillance and automated decision-making that workers and unions have flagged as their biggest immediate fears about AI on the job, and they place procedural limits on several classes of workplace AI.
Key takeaways
- Governor Gavin Newsom signed a suite of laws that ban relying solely on AI to fire workers, prohibit predicting employees’ emotional states and forbid collecting neural data.
- Under these statutes, employers must tell employees when layoffs were prompted by AI, and they may not deploy AI systems to monitor restrooms at the workplace.
- Lorena Gonzalez, president of the California Federation of Labor, AFL-CIO, called the measures “a turning point” for worker protections against AI.
- Robin Feldman of the AI Law & Innovation Institute at The University of California College of the Law, San Francisco, noted the laws offer government enforcement only, meaning workers cannot bring private lawsuits under the statutes.
Table of contents
Exactly what the laws stop and require
The new California statutes set out clear prohibitions and a notification duty intended to limit automated and invasive uses of AI at work. They ban employers from making sole decisions to terminate employment using AI, bar using AI systems to infer or predict workers’ emotional states, and forbid collecting neural data — electrical signals from the brain or nerves — for workplace purposes. The laws also outlaw AI-driven surveillance in workplace bathrooms and require employers to inform workers when AI caused layoffs. These measures are drafted to address the two core mechanisms workers fear most: automated decision- making without human review, and pervasive biometric or behavioural surveillance.
The legislation treats several categories of AI use differently: decisions that substitute entirely for a human judgement are restricted, predictive inferences about subjective states are off-limits, and certain biometric classes such as neural signals receive an explicit ban. The requirement to notify affected employees about AI-caused layoffs creates a transparency duty that advocates hope will curb opaque adoption of replacement technologies and allow collective bargaining or regulatory scrutiny to follow.
Why unions and advocates pushed for these rules
California unions and worker advocates framed the laws as a direct response to real complaints and emerging products. Lorena Gonzalez, president of the California Federation of Labor, AFL-CIO, who helped lawmakers draft parts of the package, said the measures show workers don’t have to accept automated displacement and intrusive surveillance. Union campaigns tracked products being sold and flagged likely workplace uses as a way to stay ahead of adoption: if a tool exists for sale, the federation treated that as evidence it could be deployed at scale.
Advocates pointed to specific company practices to make their case: Amazon warehouse workers’ complaints about timed bathroom breaks and reports that some healthcare employers used automated tools to rate nurses’ patient interactions were cited as examples of how surveillance and automated scoring already affect employees. The law package is therefore both a reaction to documented practices and a pre-emptive move against categories of future uses advocates expect to spread.
Limits in the statutes and how employers are responding
Legal experts and employment lawyers say the laws mark a shift but also have practical limits. Robin Feldman, director and founder of the AI Law & Innovation Institute at The University of California College of the Law, San Francisco, pointed out that the statutes include no private right of action—only government agencies may enforce them—so workers cannot sue under these specific rules. Employment lawyer Danielle Ochs of Ogletree Deakins’ San Francisco office warned employers are more focused on how to implement AI responsibly; she also said opponents worry the rules could unintentionally bar helpful safety tools if they are written too rigidly.
Some companies have already scaled back or paused internal surveillance programs after worker pushback: Meta paused a program that tracked employees’ computer activity for AI training, and other high-profile disputes have prompted lawsuits alleging biased layoff targeting. Employers will now need to weigh compliance with the new requirements against operational uses of AI, and many are likely to seek guidance on how to adapt tools so they do not trigger the statutes’ bans or notification duties.
How the laws could shape the field
The case for
- The laws may force greater transparency: the AI layoff-notice requirement creates a new trigger for regulatory review and union bargaining over automation.
- California’s approach could set a model other states adopt, as lawmakers in Colorado, Connecticut, Illinois and Texas have already passed narrower workplace AI rules.
The case against
- Without private enforcement, worker remedies may be limited to government action, which could slow or narrow practical accountability.
- Broad or rigid language could deter employers from using beneficial safety or productivity tools until regulators issue clarifying guidance.
What to be careful about
- The statutes contain no private right of action, which could leave affected employees dependent on government enforcement and delay remedies.
- Overbroad drafting risks banning legitimate safety uses of AI if regulators do not provide clear implementation guidance.
The bottom line
California’s package is notable for its direct limits on several classes of workplace AI and for forcing disclosure when AI causes layoffs. The statutes respond to documented complaints about surveillance and scoring at employers such as warehouses and healthcare providers, and they give unions and regulators new levers to contest deployments. At the same time, the lack of private enforcement and the potential for broad drafting mean much will depend on how agencies interpret and implement the rules and whether other states adopt similar models.
What to watch
- Watch whether California or state agencies publish enforcement guidance or interpretive rules; no date has been set.
- Watch for renewed legislation from the California Federation of Labor to require employer disclosure of workplace AI tools; no date has been set.
- Watch other state legislatures for bills that copy or narrow California’s approach, including Colorado, Connecticut, Illinois and Texas; no dates have been set.
Frequently asked questions
What do California’s new AI laws ban?
The statutes ban employers from relying solely on AI to fire workers, bar using AI to predict employees’ emotional states, forbid collecting neural data and prohibit AI surveillance in workplace bathrooms.
Who can enforce the new rules?
Robin Feldman of the AI Law & Innovation Institute at The University of California College of the Law, San Francisco, noted the laws provide government enforcement only, meaning workers cannot bring private lawsuits under these specific statutes.
Will the laws prevent all job losses to AI?
Advocates say the laws create procedural limits and transparency that can slow or reshape automated replacements, but employment lawyer Danielle Ochs warned it is unclear how sweeping the change will be because the statutes may not address all real-world employer uses of AI.
Related reading
