California's New AI Health Care Law Faces Sacramento Strike Reports

Governor Gavin Newsom signed AB 1979 on September 30, 2026, and its health care rules take effect January 1, 2027. The law says AI cannot independently perform clinical work reserved for licensed professionals in California health facilities. On Sunday, November 1, a report began circulating that Sacramento healthcare workers had struck over AI protections, roughly two months before those rules start to apply.

That strike report is the shakiest part of this story. It comes from a Substack roundup, Rio Rundown, which cited Hoodline. We could not confirm the walkout in a primary or local outlet, so treat its details as reported rather than established. The statute itself and Kaiser Permanente's labor fight earlier this year stand on firmer ground.

What California's AI health care law requires starting January 1, 2027

AB 1979 was sponsored by the California Nurses Association and authored by Assemblymember Mia Bonta. It bars health facilities from letting AI independently perform clinical work reserved for licensed professionals. It also requires them to keep clinicians' independent judgment in place when AI-supported clinical decision tools inform care, according to LexCircuit and Governing.

The law does not ban AI from clinics. It limits what those tools can do on their own, and it asks health systems and clinics to take reasonable steps to keep licensed clinicians in charge of AI-informed care from January 1, 2027.

The law also brings health chatbots under medical privacy rules. Businesses offering health chatbots to California consumers become subject to the state's medical confidentiality rules, as Healthcare IT News reported.

The statute leaves a key term undefined. It does not define "independent professional judgment," a gap that Fox Rothschild's privacy blog flags. The same post notes a similar gap in California's new law barring lawyers from delegating legal work to generative AI. How the state will define the term is still undecided, so early enforcement disputes will do much of the defining.

Newsom framed the signings around workers. In a statement, he said "AI should expand opportunity — not come at the expense of workers and families." Sandy Reding, RN and president of the California Nurses Association, said in a press release that AI "must support, not replace, the clinical judgement of the nurses."

Hospital groups and the California Medical Association took neutral positions on AB 1979, while a hospital association spokesperson said clinicians should keep final authority over care. That split suggests the cost of compliance will fall most heavily on the organizations that buy and configure clinical software.

Why Newsom vetoed the nurse override bill

The same day he signed AB 1979, Newsom vetoed AB 2575. That bill would have shielded nurses from retaliation for following or overriding an AI tool. Nurse.org reported the veto alongside the clinical judgment law he did sign.

Newsom's reasoning centered on who decides. He said the bill would require the Labor Commissioner to judge an override against a standard, "something the Labor Commissioner does not have the medical expertise or skill to do."

The practical result is narrower than the bill's supporters wanted. Nurses and therapists now have a legal basis to argue that AI cannot replace their judgment. They do not have explicit protection for overriding an AI tool at work. That gap is likely to show up at the bargaining table, where unions can negotiate override protections the statute does not provide.

Sacramento strike reports: what is and is not confirmed

The Rio Rundown roundup says Sacramento healthcare workers struck over AI protections shortly after state lawmakers enacted rules on workplace AI. It also says some nurses saw sentiment scores in call metadata, though the employer denies using such tools.

Those claims need checking on four fronts: the date, the number of participants, the employer involved, and the call-metadata allegation. A roundup that cites another outlet is a lead, not a primary record. Until primary or local reporting confirms the details, readers should not assume the walkout's size or scope.

Kaiser's March strikes set the labor template

Kaiser Permanente's therapists and nurses staged AI-related strikes earlier in 2026, in March, covering Sacramento and other Northern California facilities. The sources do not agree on the details.

Kaiser's therapists have worked without a contract since September 2025, according to National Today. A late-2025 stoppage involved up to 31,000 health workers across California and Hawaii, per LawSnap.

LawSnap's legal tracker calls the Kaiser dispute one of the first major labor actions centered on AI integration rather than wages or staffing. That framing matters. It moves AI from a side issue in contract talks to a central demand, and it gives other health systems a template to watch.

Kaiser has disputed the emotion-reading part of the claims. A company spokesperson, whose name the source did not give, told KQED: "We currently do not use AI tools to recognize, predict or infer emotional states of our nurses." Kaiser also said it is reviewing what AB 1883 means for its operations.

Emotion-reading AI and California's bossware fight

AB 1883 restricts employers' use of AI to read workers' emotions. KQED reports that California has gone further than any other state to regulate "bossware," the term for AI surveillance software used on workers.

Call centers are the obvious test case. Sentiment scoring of calls, the kind of tool the Sacramento report describes, looks like the sort of emotion inference the law targets, though how far the statute reaches will depend on enforcement. Employers and vendors selling workplace monitoring tools now have to plan for those limits, KQED's reporting suggests.

The federal backdrop matters too. As of this writing, the United States has no comprehensive federal AI law, which leaves states to set rules in areas like health care and employment. Federal preemption could still override state rules, but reporting for this story did not identify a current federal move.

What health systems, chatbot users and workers can do now

No state agency has published step-by-step guidance on these issues that we could find, so the steps below are general advice. Health chatbots are also no longer a niche product. Surveys covered by USA 24 found that two-thirds of tweens and teens use AI chatbots, which makes the privacy question broad.

• If you use a health chatbot in California, read its privacy policy and check how it handles your medical records and whether it says it follows state medical privacy rules.
• If you are a patient, ask your provider whether AI tools inform your care and who makes the final clinical call.
• If you are a worker, ask your employer whether it uses AI tools to monitor, score or summarize your calls or sessions.
• Ask directly whether any tool infers emotional state or sentiment, and request the answer in writing.
• If you run a clinic or health system, list every AI tool that informs clinical decisions and record which licensed clinician reviews each one before January 1, 2027.
• Raise these questions with your union or professional association before contract talks or grievance deadlines.

What to watch as AB 1979 takes effect

The next test is January 1, 2027, when the health care provisions start. Three things bear watching. First, how the state defines "independent professional judgment," and whether the Labor Commissioner or the courts will enforce these provisions, which remains unclear. Second, whether Kaiser's review of AB 1883 produces policy changes that other employers copy. Third, whether primary or local reporting confirms the Sacramento walkout.

Regulated industries are already drawing hard lines around AI agents. Five major U.S. banks have blocked outside AI agents from customer accounts, as USA 24 reported. Employers that treat California's rules as a compliance checkbox may find that the harder questions arrive in grievances and contract disputes, not in the statute itself.