Mental health workers in California have done something unusual. They’ve gone on strike — not over pay, not over hours, but over artificial intelligence.
More than 2,000 therapists, psychologists, social workers, and other behavioral health professionals at Kaiser Permanente walked off the job this week in what their union is calling the first mental health strike in the United States driven primarily by concerns over AI replacing human clinical judgment. The work stoppage, organized by the National Union of Healthcare Workers (NUHW), began Monday and is expected to last five days, affecting tens of thousands of patients across Northern California, as reported by Futurism.
The core grievance is blunt: Kaiser Permanente has been deploying AI-powered tools that workers say are being used to screen patients, recommend treatment plans, and even determine the number of therapy sessions a person should receive. Clinicians argue these systems are making decisions that should belong to trained professionals who can read a room, interpret silence, and understand what a patient isn’t saying.
“We didn’t go to school for a decade to have an algorithm tell us how to treat a suicidal teenager,” one striking therapist told reporters outside a Kaiser facility in Oakland.
That sentiment runs deep.
Kaiser Permanente, the largest managed care organization in the country with over 12.7 million members, has invested heavily in digital health infrastructure. The company has publicly touted its use of technology to improve access to mental health care, a sector facing an acute labor shortage nationwide. The Bureau of Health Workforce projects a shortfall of tens of thousands of mental health professionals by 2030, and Kaiser has positioned AI as part of the answer to that gap.
But the workers on the picket line see it differently. They contend that AI tools are being used not to supplement care, but to ration it. According to NUHW representatives, Kaiser’s AI systems have been involved in triaging patients who call seeking mental health support, sorting them into categories that determine what level of care — if any — they receive. Some patients who might benefit from ongoing psychotherapy are instead routed to app-based self-help programs or group sessions, clinicians say.
The implications extend far beyond one health system in Northern California. Mental health care in the United States is already stretched to its breaking point. The National Alliance on Mental Illness reports that over 60% of U.S. counties have no practicing psychiatrist. Wait times for an initial therapy appointment routinely exceed six weeks in major metropolitan areas. And the post-pandemic surge in anxiety, depression, and substance use disorders has only intensified demand.
So when a system as large as Kaiser starts deploying AI to manage patient flow in behavioral health, the rest of the industry pays attention.
The technology itself isn’t inherently controversial. AI-assisted tools in health care have shown promise in radiology, pathology, and even some areas of primary care diagnostics. The difference, striking workers argue, is that mental health care depends on the therapeutic relationship — a bond between clinician and patient that is, by definition, human. A radiologist reviewing an AI-flagged scan still makes the final call on a diagnosis. But when an AI system determines that a patient’s self-reported symptoms don’t warrant individual therapy, the clinician may never meet that patient at all.
That’s what worries people like Dr. Ravi Mehta, a clinical psychologist and NUHW member, who told Futurism that the tools are “making clinical decisions without clinical context.” He described cases where patients with complex trauma histories were assessed by AI screening questionnaires and categorized as low-risk, only to later present in crisis.
Kaiser Permanente has pushed back against these characterizations. In a statement, the organization said its AI tools are designed to “enhance clinician decision-making, not replace it,” and that all treatment decisions ultimately rest with licensed providers. The company also pointed to its significant investments in hiring — Kaiser added thousands of mental health clinicians in recent years following a 2023 settlement with California regulators over long wait times for therapy appointments.
That settlement matters. In 2023, Kaiser agreed to pay $200 million in penalties and remediation after the California Department of Managed Health Care found systematic failures in providing timely mental health care. The company promised to hire more staff and reduce wait times. Workers now say the AI push is Kaiser’s way of meeting those mandates on the cheap — using technology to process more patients without proportionally increasing the clinical workforce.
It’s a familiar tension in American health care. Technology promises efficiency. Workers see cost-cutting. Patients get caught in the middle.
The strike has drawn support from surprising corners. Several California state legislators have issued statements backing the workers, and Assemblymember Matt Haney, who has been active on AI regulation, called the situation “a canary in the coal mine for every profession that requires human judgment.” The California Psychological Association, while not officially endorsing the strike, released a statement expressing “serious concern” about the use of AI in clinical mental health settings without adequate oversight or transparency.
And the timing is notable. The strike comes as federal and state regulators are grappling with how to govern AI in health care more broadly. The Biden administration’s 2023 executive order on AI safety included provisions for health care applications, and the FDA has been slowly building a framework for regulating AI-enabled medical devices. But mental health screening tools occupy a gray area — they often don’t meet the threshold for FDA device regulation, yet they can profoundly influence patient care pathways.
This regulatory gap is precisely what concerns labor advocates. Without clear rules about when and how AI can be used in clinical mental health decisions, health systems have wide latitude to deploy these tools as they see fit. Workers have little recourse beyond collective action.
Which is exactly what they’ve chosen.
The NUHW has outlined several demands. They want Kaiser to disclose which AI tools are being used in behavioral health, how they function, and what data they rely on. They want clinicians to have the right to override AI recommendations without managerial pushback. And they want a moratorium on expanding AI-driven triage in mental health until an independent review can assess the tools’ accuracy and impact on patient outcomes.
These demands mirror a broader pattern emerging across industries. Writers, actors, dock workers, and now health care professionals — all have drawn lines around AI in recent labor actions. But the mental health context adds a layer of urgency that’s hard to dismiss. A chatbot that writes a mediocre screenplay is one thing. An algorithm that decides whether a person in psychological distress gets to talk to a human being is something else entirely.
Research on AI in mental health settings is still in early stages, and the results are mixed. A 2024 study published in JAMA Psychiatry found that AI-based screening tools correctly identified major depressive disorder about 80% of the time — impressive, but still wrong one in five cases. Other studies have raised concerns about algorithmic bias, noting that AI models trained on predominantly white patient populations may underperform when assessing symptoms in Black, Latino, and Asian patients, populations already underserved by the mental health system.
Kaiser’s patient population is one of the most diverse in the country. The equity implications are not theoretical.
On the picket lines, workers have been joined by patients and their families. Several current Kaiser members told local media they had experienced what they described as impersonal, automated interactions when seeking mental health support. One woman said she completed an online assessment, was told her symptoms were “moderate,” and was directed to a digital wellness program. She later attempted suicide. She’s now in treatment with a human therapist — one she had to fight to see.
Stories like hers are anecdotal, and Kaiser would likely argue they don’t represent the system’s overall performance. But they resonate because they confirm a fear that many Americans share: that the health care system values throughput over care, and that AI is accelerating that tendency.
The financial incentives are real. Mental health care is expensive. Therapists are expensive. AI screening and triage tools are, comparatively, not. For a system like Kaiser that operates on a capitated model — receiving a fixed payment per member per month regardless of how much care is delivered — anything that reduces the cost of care delivery directly improves the bottom line. Workers understand this arithmetic, and they resent it.
Kaiser reported $3.3 billion in net income in 2023. The organization is not struggling financially. That fact has not been lost on striking workers, who see the AI push as a choice driven by margin optimization rather than clinical necessity.
The five-day strike will inevitably disrupt care for patients who are already waiting too long. Kaiser has said it will use management staff and temporary workers to cover essential services during the walkout. But mental health care isn’t like emergency medicine — you can’t just swap in a replacement. The therapeutic relationship depends on continuity, trust, and familiarity. Every disruption sets treatment back.
Workers know this. They say they struck reluctantly, and only after months of failed negotiations.
What happens next will depend on whether Kaiser comes to the table with meaningful concessions on AI transparency and clinician autonomy. The NUHW has signaled willingness to negotiate but says it won’t accept vague assurances. They want enforceable contract language.
The broader question is whether this strike becomes an isolated incident or a template. Mental health professionals across the country are watching. So are nurses, physicians, and other clinicians who increasingly find AI tools embedded in their workflows — sometimes without their input or consent. If Kaiser’s workers win concrete protections, it could set a precedent that reverberates through health care labor negotiations for years.
And if they don’t, the message will be equally clear: that in the contest between algorithmic efficiency and human clinical judgment, the algorithm has the upper hand.
For now, on the sidewalks outside Kaiser facilities in Oakland, San Francisco, Sacramento, and San Jose, therapists hold signs that read “Patients Need People, Not Programs” and “AI Can’t Hold Space.” The phrases sound like slogans. They’re also, according to the people holding them, statements of clinical fact.
The consulting room has always been one of the last places where being human was the whole point. Whether it stays that way may depend on what happens this week in California.
WebProNews is an iEntry Publication