# California’s New AI ‘Bossware’ Laws Are About to Take Effect. Are Employers Ready? > Source: > Published: 2026-10-28 14:00:00+00:00 # California’s New AI ‘Bossware’ Laws Are About to Take Effect. Are Employers Ready? Pa Vue is a telephone triage nurse with [Kaiser Permanente](https://ww2.kqed.org/news/tag/kaiser-permanente). She works the night shift, from midnight to 8:30 a.m., from her home in the Central Valley. Vue’s job is to calmly assess each caller as they describe their concerns and symptoms, looking for red flags that indicate immediate treatment is warranted, like signs of a heart attack or stroke. It’s a critical job, and one of a growing number impacted by AI-powered surveillance software, aka “bossware.” California has gone further than any other state to begin regulating it. At the end of September, Gov. Gavin Newsom signed [five labor-backed bills regulating AI in the workplace](https://ww2.kqed.org/news/12102337/newsom-signs-slate-of-ai-workplace-laws-barring-robo-bosses-and-surveillance). Among them is [AB 1979](https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260AB1979), which bars health facilities from letting AI alone do work reserved for licensed professionals, and requires them to protect clinicians’ independent judgment when AI informs care. [AB 1883](https://leginfo.legislature.ca.gov/faces/billVersionsCompareClient.xhtml?bill_id=202520260AB1883) restricts employers from using AI-powered workplace surveillance tools to read workers’ emotions or collect their neural data. In 2024, Vue said she learned Kaiser was piloting a feature that included scoring the sentiment of recorded calls — as in, evaluating employees’ attitudes in the recordings. Nurses were later told the pilot was paused. But while Vue was on maternity leave a year later, Kaiser appeared to have instituted the feature, because Vue said she came back to work to find a “sentiment score” in the metadata attached to each of her recorded calls. Vue said she isn’t opposed to AI in medicine, but she objects to the sentiment scoring because triage nurses are trained to stay calm and press patients to detail their symptoms, even if that means the patients become agitated or fearful in the process. Vue said she wasn’t even sure how to interpret the numbers she was seeing in the metadata. She wasn’t sure what scale she was being graded on or how high the numbers went. The nurses asked management about the metrics, Vue said, but got no answers. Whoever made the system that was evaluating the calls, Vue said, didn’t “really understand what we’re going through.” Contacted for comment, a spokesperson for Kaiser wrote KQED, “We currently do not use AI tools to recognize, predict or infer emotional states of our nurses.” The spokesperson also wrote that the healthcare giant is reviewing the implications of AB 1883 for its operations. “Scoring whether someone said the right words is one thing, and that’s okay, but scoring what they were feeling now is off limits,” management-side employment attorney Anthony Zaller told KQED. AB 1883 should, in theory, require removing sentiment scoring-style software from jobs like Vue’s. Zaller said he’s not familiar with Kaiser’s use of AI specifically, but he said that employers commonly don’t realize the extent to which AI is embedded in their software — or that existing laws already make employers responsible in a way they can’t sidestep “just because you ran it through some software.” Vue is a member of the California Nurses Association, which has been [battling Kaiser](https://ww2.kqed.org/news/12097878/in-oakland-kaiser-nurses-take-staffing-ai-concerns-to-the-picket-lines) over its use of AI bossware. Her account appeared under a pseudonym as one first-person example among a couple dozen from workers around the state in a recent labor coalition [report](https://url.us.m.mimecastprotect.com/s/wjHhCwpkPvsknn3yFVfGpsJn6NG?domain=report.norobobosses.org/) warning that artificial intelligence and surveillance tools are increasingly used to monitor, manage and discipline workers. “Workers are experiencing unprecedented levels of surveillance and monitoring in their workplaces,” said Stasia Hansen, deputy director of research and strategy at one of the coalition members, the East Bay Alliance for a Sustainable Economy, a labor and housing organization focused on Alameda and Contra Costa counties. Vue first learned of the 2024 pilot because she sits on a joint union-management committee on nursing practice; the committee brought its concerns to management’s attention “to really hold them accountable to their own principles,” she said, referring to [Kaiser’s publicly stated AI policy](https://about.kaiserpermanente.org/expertise-and-impact/public-policy/our-key-issues/artificial-intelligence). She was less than impressed with management’s response. “They tell us they can’t turn it off. Of course, nothing can be turned off because it’s built into the system. So they will have to lose a lot of money and go with a different software, or they’re just going to have to use the platform,” Vue said. KQED asked Kaiser whether its call platform generates sentiment scores, whose sentiment they measure and whether the feature can be turned off. Kaiser did not answer. How widespread bossware is remains hard to measure. Employers rarely disclose the tools they use, and the algorithms are proprietary, so outside researchers seldom get to examine them except through leaks or lawsuits. But the signs keep piling up. The use of work-surveillance technology took off during the pandemic. As many people started working remotely, more employers began surveilling their teams in new ways: tracking keyboard strokes, taking screenshots and monitoring pauses. According to The New York Times, by 2022, [eight of the 10 biggest private employers](https://www.nytimes.com/interactive/2022/08/14/business/worker-productivity-tracking.html) in the United States were tracking the productivity metrics of individual workers. Workers in a wide variety of industries, Hansen said, “are facing consequences due to productivity expectations or due to systems that don’t reasonably take into consideration key factors of their workplaces and work situations, which we as humans are able to assess.” As it moved through the legislature last session, [AB 1883](https://ww2.kqed.org/news/12097632/chatbots-data-centers-and-surveillance-5-silicon-valley-bills-land-on-newsoms-desk) faced opposition among business-friendly interests. The California Chamber of Commerce, for example, said in a letter to lawmakers that the bill would open employers up to litigation over what the phrase “emotional state” means. The Chamber also said that the ban could sweep in AI security systems that detect agitation or aggression, tools some employers adopted to comply with a 2023 state law requiring workplace violence prevention plans. California lawmakers aren’t the only ones to consider legislating limits on bossware. The European Union, for instance, has banned emotion recognition at work since 2025. But similar bills specifically seeking to regulate the use of AI to detect worker emotion died this year in Minnesota, New York and Washington. A broader Massachusetts bill limiting workplace monitoring stalled in committee. California’s bills “took some hits in the committee process,” said Hansen, the East Bay Alliance for a Sustainable Economy deputy research director, referring to the compromises made to get the proposals through the legislature and signed by the governor. What can pass in Sacramento, she said, is different from what can pass locally. “So while we’re thrilled to see this foundation at the state level, we also know that there’s a lot more to be done for actual robust protections for workers, and so we’re looking to move those on the local level, across the state.” Hansen did not name a Bay Area city or county that has agreed to take up such a measure. None appears to have done so, although Santa Clara County supervisors voted in March to study AI use across county departments, and San José’s largest city employee union has sought limits on AI in contract talks with the city. The East Bay coalition, which includes the San José nonprofit Working Partnerships USA, held its launch at San José City Hall in September. Hansen said the coalition plans to start with public employers, because cities can act both as employers — setting rules for their own workforce and the technology they buy — and as regulators of private employers within their limits. “It’s both the city as an employer and also the city as a regulatory body,” she said. In the meantime, tech companies are pouring some of the largest private investments in history into AI built to automate work. Zaller, the employment attorney, is doubtful regulation will stop the use of bossware altogether. “The Pandora’s box is open. It’s here,” he said.