On AI, Newsom gives labor most of what it demanded

On the final day to act on his last scheduled legislative session, Gov. Gavin Newsom on Wednesday signed several bills to protect workers from artificial intelligence, including limits on the use of AI in patient care and in firing people.

At the same time, the governor vetoed two AI-related bills labor leaders had pushed for, which would have protected health care workers.

Enacting the labor bills will not only extend new protections to workers but may also improve Newsom’s odds in his expected bid for the presidency. Earlier this month, Newsom traveled to South Carolina, the first state to vote in the 2028 Democratic primary.

Labor leaders have pushed hard for the governor to sign a slate of nine AI-related bills, saying it would affect their support for any presidential run. The slate includes bills cosponsored by the California Labor Federation, Communication Workers of America, SAG-AFTRA, Teamsters and the California Nurses Association. Last year, Newsom vetoed several AI bills backed by labor unions

Unions, having won on seven of their nine bills, were generally supportive of the governor.

“We’re glad Governor Newsom is catching up with the overwhelming sentiment of the American people on AI and starting to lead on this issue,” California Labor Federation president Lorena Gonzalez told CalMatters in a written statement. “We still have work to do, especially to regulate AI in healthcare and demand human oversight of these technologies.”

California Nurses Association President Sandy Reding wrote in a statement that Newsom’s signing of the labor bills made this “a historic day in California.” 

Nicolle Pillion, who answers patient phone calls at Kaiser Permanence, said she has struggled with technology designed to score her empathy because she has a hoarse voice. She called Newsom’s moves “a big win for us and a step in the right direction” even though she’s not happy with the vetoes.

AFL-CIO spokesperson Steve Smith said it’s too soon to discuss an endorsement, but in a statement provided to CalMatters, AFL-CIO president Liz Shuler called the new laws “some of the strongest state legislation in the country to protect working people from the harms of unregulated, untested AI” and a sharp contrast to a voluntary deal Big Tech companies and President Trump agreed to yesterday.

The bills Newsom signed into law Wednesday will:

• Limit the use of AI in terminating or disciplining workers
• Ensure that health care professionals, not AI, have the final say in decisions about patient care
• Prohibit employers from making predictions about a worker’s emotional state and the collection of brain data
• Prevent employers from using AI-powered surveillance in bathrooms
• Require employers to tell workers if they’re getting laid off in whole or in part because of AI or automation

Those he vetoed would have:

• Protected health care workers from retaliation if they turn down a recommendation by a clinical decisionmaking system
• Limited the way AI is used in mental health settings

In veto messages, Newsom called the mental health worker law “overly broad” and said he struck down the other measure in part because  the Labor Commissioner lacks the medical expertise to enforce the bill’s anti-retaliation rules.

Reding of the nurse’s union criticized the veto of the anti-retaliation bill, saying, “he vetoed the bill that would have protected us for using our judgment,” and “that makes no sense to any nurse at the bedside. Today, a nurse who sees that the algorithm is wrong and acts to save her patient can still be disciplined or fired by her employer.”

Newsom enacted additional employee AI protections earlier this month, including a law that keeps K-12 and Cal State University schools from replacing faculty with AI, and laws that make advertisers disclose when they depict AI-generated performers and make businesses provide people with a human customer service agent within 15 minutes.

The new measures build on other legislation Newsom approved to rein in AI. Since the California legislative session came to an end one month ago, Newsom signed several bills to regulate the technology and related systems, including several measures to curb data centers’ utility costs and environmental harms, more than a dozen bills to protect kids from technology, and a pair of bills that foster more monitoring of AI models.

The law curbing AI in firing or disciplining workers has been weakened throughout the legislative process, losing an appeals process for workers, the right to bring a lawsuit to compel an employer to follow the law, and it no longer covers contractors.

“It should be a stronger bill, but we absolutely believe the American people believe, Californians believe, that an individual should not have the indignity of being fired or disciplined by a robot, by an algorithm, without human oversight,” Gonzalez said earlier this month.

Concerns about AI doom have gripped headlines in recent weeks, fueled by a series of incidents in which AI systems hacked into online computer systems, but a Pew Research poll shows that fear of AI job loss is the most significant reason people in the U.S. say they’re concerned about AI. The stakes around AI safety have also grown as AI is increasingly used to manage or evaluate workers and decide how much a worker gets paid.

In a letter to Newsom earlier this month, AFL-CIO president Liz Shuler called the slate of AI bills “a floor to protect workers from the worst abuse of AI in the workplace while allowing for adjustments as technology advances.”

As AI increasingly enters the workplace, unions have pushed for worker protections in the agreements they negotiate with employers and used the issue to stoke membership. Kaiser Permanente nurses preparing for contract negotiations recently told CalMatters they’re concerned about AI and surveillance getting in the way of their duty of care to patients. Unions also supported a bill restricting the use of brain data amid fears that employers will collect and misuse such data in the future.

Gonzalez said the laws signed by Newsom are necessary not just because use of AI in the workplace is increasing but because the majority of people aren’t part of a union and therefore do not take part in collective bargaining over how their employers use the technology . More than 70 percent of people in the U.S. approve of unions, according to a Gallup poll released earlier this month, a level of approval not seen since the 1960s, yet elections to recognize unions, which require approval by a federal agency, fell 30 percent last year.

At a rally and press conference earlier this month in San Francisco, Gonzales detailed the kind of laws the labor federation wants to see get enacted next year, including a prohibition on companies forcing workers to train the AI that replaces them and renewed pushes for legislation that failed previously, including a ban on surveillance pricing and a requirement that a human operator be present in autonomous big rig trucks.

A majority of Americans, and for the first time a majority of people under 30, say they are more concerned than excited about the future of AI, according to a Pew Research survey conducted in June. Closer to home, roughly half of Californians say AI is used in their workplace extensively or moderately, and two out of three Californians want the government to adopt policy to protect them from job loss, according to a Carnegie Endowment California survey conducted this spring.

Fear of AI job cuts led Newsom to sign an executive order earlier this year that orders state employees to study the impact of AI on job loss and explore options like job training , stock compensation, and cooperative business ownership for workers.

Newsom’s time as governor coincided with a boom in AI investment and use. He took office the same year that Google released BERT, an open source language model that pioneered techniques to predict the next word in a sequence or generate text later used by popular systems including OpenAI’s ChatGPTin 2022.

Newsom has walked a line between enacting some regulations in the name of safety and blocking others in the name of protecting innovation and competitiveness. In fall 2022, Newsom  ordered researchers and state agencies to explore the opportunities and risks tied to generative AI, pushing state agencies to launch generative AI pilot projects to tackle problems ranging from taxes to traffic to homelessness. Two years ago and again this year Newsom ordered state agencies to consider safety and potential harm before signing contracts. In 2024 he vetoed an ambitious effort to mandate testing of major AI models for catastrophic risks. This year, Newsom at least partly reversed himself on that veto, signing an executive order to fast track a law that empowers independent organizations to verify claims made about AI models and to explore the possibility of a “kill switch” to shut down AI models in an emergency.