California will make employers flag AI-driven layoffs and curb algorithm firings

Starting Jan. 1, 2027, a California employer whose mass layoff, relocation or termination is caused "in whole or in substantial part" by AI or other automation will have to say so at the top of its WARN notice, in a prescribed sentence: "This notice is for a technology displacement." The notice must also give the number and type of positions affected, the job functions being automated and the category of technology responsible. The state will publish a quarterly summary of those filings, which employment law firm Fisher Phillips warned employers could draw attention from the public, unions and plaintiffs' attorneys.

That requirement, in SB 951, is one of four workplace AI bills Gov. Gavin Newsom signed on Sept. 30, KQED reported. The others put a human reviewer between automated systems and disciplinary or firing decisions, ban AI tools that read workers' emotions or collect neural data, and bar surveillance in workplace bathrooms. Newsom's office billed them as first-in-the-nation protections. SB 951 also directs the state Employment Development Department to study AI's effect on hiring and the workforce.

The package reaches far into how workers are managed and let go. The headline law stops at the front door. SB 947, the "No Robo Bosses Act," applies only to current employees, not job applicants, according to Fisher Phillips. AI in hiring is governed instead by regulations from the California Privacy Protection Agency, effective Jan. 1, 2027, which require employers using AI and similar tools for hiring and other employment or compensation matters to issue pre-use notices, offer individuals the chance to opt out and complete risk assessments, HR Dive reported.

Under SB 947, employers may not rely solely on an automated decision system to discipline or fire someone. When an employer "primarily relies" on one, a human reviewer must corroborate the output using supervisory evaluations, personnel files, work product or similar evidence. The employee must receive a standalone written notice when the decision is delivered and can request a description of their own data the system used over the past 12 months, Fisher Phillips said. Affected workers must also be given a human point of contact who can explain the decision, according to CNBC.

The law also bars using an automated system to infer a worker's protected characteristics or to predict and retaliate against a worker for exercising legal rights, attorneys at Ogletree Deakins wrote. Violations carry a $500 penalty each. The law takes effect July 1, 2027.

SB 947 is the narrower version of the bill. Newsom vetoed its predecessor last year, writing that "rather than addressing the specific ways employers misuse this technology, the bill imposes unfocused notification requirements on any business using even the most innocuous tools." When state Sen. Jerry McNerney reintroduced it in February, he dropped the advance-notice requirement and stripped language that would have extended its protections to gig workers, which had drawn heavy criticism from Uber and Lyft, CNBC reported. Rideshare drivers, classified as independent contractors under Proposition 22, are outside the law. "An algorithm decides what work we get, what we are paid, how our performance is judged, and too often whether we can work at all," Nicole Moore, president of Rideshare Drivers United, wrote, according to KQED.

The trigger for most of SB 947's obligations is still contested. In a letter urging a veto, Robert Singleton of the Chamber of Progress argued that "primarily relies" is never defined. "Employers are given no objective standard for determining when a technology has moved from merely informing a decision to being a primary basis for it," he wrote. Jackson Lewis attorneys advised employers to inventory workplace tools that may fall under the state's definition of an automated decision system and to evaluate how their outputs could affect covered decisions.

The surveillance laws take effect sooner, on Jan. 1, 2027, and may catch software employers do not think of as surveillance. AB 1883 bans AI-powered workplace surveillance tools that recognize or predict an employee's emotional state or collect neural data, with an exception for tools used to ensure safety. Fisher Phillips said the ban likely reaches sentiment and mood features in contact center, collaboration and HR analytics platforms, including tools that analyze emails and chats. AB 1331 defines a surveillance tool broadly enough to cover access badges, wearables, geolocation on company phones and time-tracking software that logs idle time, the firm said. Employees may set such devices down before entering a bathroom without discipline, and employers may not use restroom access data to track how often or how long someone is inside. Both laws carry penalties of up to $500 per violation. The statutes do not give workers an express right to sue, but Fisher Phillips and Ogletree both flagged potential exposure under the Private Attorneys General Act.

The tools these laws target are already widespread. In an OECD survey cited by CNBC, 90% of U.S. managers said their firms had adopted at least one tool to "instruct, monitor or evaluate workers," the highest share of any country surveyed. Meta faces a lawsuit, filed in July, in which former employees allege AI-assisted systems were used to rank and select workers for layoffs. Meta has denied the allegations.

Newsom did not sign everything put in front of him. On the same day, he vetoed AB 2656, which would have required notice to public employees when AI performs work within the scope of their jobs, KQED reported. California will now require employers to announce when automation eliminates jobs in bulk. It will not require them to tell public employees when AI is doing part of their work.