Workers affected by those decisions must also receive written notice explaining that AI was used as a primary factor. Employers must disclose what employee data the system considered and provide a human contact who can explain the decision further.
“No worker should ever be fired or disciplined by a machine, AI or not. Artificial intelligence systems have the potential to boost productivity, but they’ve also made errors and misjudgments and exhibited bias,” said California State Senator Jerry McNerney, who authored the bill. “AI must remain a tool controlled by humans, not the other way around.”
The law marks a return for legislation Newsom rejected in 2025. That earlier version would have required companies to notify workers in advance when certain automated systems could affect workplace conditions. Newsom said at the time that the proposal was too broad and could apply to relatively routine tools.
McNerney reintroduced the bill after removing the advance-notification provision and language that would have extended protections to gig workers. The changes reduced some of the objections that had surrounded the previous proposal, although business groups continued to oppose the measure.
One point of criticism centered on the phrase “primarily relies,” which determines when the law's additional requirements apply. Robert Singleton, the Chamber of Progress' senior director of policy and public affairs for California and US West, argued that the bill did not clearly define when an automated system moves from assisting a decision to becoming its main basis.
“The bill’s obligations generally apply when an employer ‘primarily relies’ on an automated decision system, but that critical term is never defined,” Singleton wrote in a letter urging Newsom to veto the legislation.
Labor groups, meanwhile, backed the law as a workplace protection. Lorena Gonzalez, president of the California Federation of Labor Unions, AFL-CIO, said workers had pushed the state to set limits on how AI is used by employers.
“When working people organize, we get results. Workers across California have demanded that our state lead the way in regulating AI in our workplaces. And today, we see that begin to happen,” Gonzalez said after Newsom signed the bill.
California's action comes as employers increasingly adopt software that can monitor, evaluate or manage workers. OECD survey results cited in the article found that 90% of U.S. managers said their companies had adopted at least one tool used to instruct, monitor or evaluate employees.
Concerns over automated employment decisions have also surfaced in litigation and worker surveys. Former Meta employees filed a lawsuit in July alleging that AI-assisted systems were used to rank workers for layoffs in a way that disproportionately affected employees who had taken medical or family leave. Meta has denied the allegations.
Workers at Walmart and Amazon have also expressed concern about automated HR decisions, according to a May survey from nonprofit United for Respect. The group later failed to persuade Walmart shareholders to approve a proposal seeking greater disclosure around the company's use of AI with workers.
California is not the only state considering restrictions on automated employment systems. Illinois has already enacted a law requiring disclosure when AI is used for certain employment purposes, though that measure does not prohibit AI from being the sole decision-maker.
Similar proposals have also surfaced in New York, Louisiana and New Jersey. At the federal level, Sens. Ed Markey and Brian Schatz introduced legislation in June that would limit employers' ability to rely on automated systems for workplace decisions.
The Electronic Frontier Foundation said the California law is a “strong step toward giving workers the protections they need in workplaces that use automated decisionmaking systems” and said it plans to continue working with labor groups on workplace AI policy.
This analysis is based on reporting from CNBC.
Image courtesy of Unsplash.
This article was generated with AI assistance and reviewed for accuracy and quality.
