San Francisco's Board of Supervisors adopted a resolution Tuesday that pressures Kaiser Permanente to pull back its use of artificial intelligence in mental health care. The vote follows months of escalating tension with 2,400 Northern California behavioral health clinicians who say contract proposals would let the healthcare giant replace therapists with AI and cut them out of care decisions.
The resolution urges Kaiser to revise proposed contracts that the National Union of Healthcare Workers (NUHW) says would give the organization sole discretion to modify how mental health services are delivered, refer patients to third-party contractors, and lay off Kaiser clinicians. The union is in mediation with Kaiser after four days of talks in August failed to produce a settlement. No additional mediation dates have been set.
"Kaiser officials should be eager to work with their clinicians to improve services, but instead they want to sideline them to escape accountability for under-resourcing behavioral healthcare," NUHW president emeritus Sal Rosselli said in a union news release. "By taking this stand, San Francisco supervisors are making it clear to Kaiser that it needs to work with therapists, not replace them with AI that won't advocate for patients."
What the contract language actually says
The disputed contract terms go beyond AI replacement. According to the union, Kaiser's proposal would let the organization lay off its own clinicians while referring patients to outside contractors, and retain sole authority to determine or change how mental healthcare is delivered. The resolution calls for removing all of these provisions.
Kaiser already uses AI to record therapy sessions and runs algorithms to screen patients seeking care, the union said. The organization has declined to add language explicitly stating AI will not replace clinicians - a commitment it made in its 2024 contract with Southern California clinicians.
In an email statement Wednesday, Kaiser said its screening tool "does not use AI to diagnose patients, make clinical decisions, or determine medical necessity. Clinical decisions remain with licensed health care professionals."
A pattern of regulatory scrutiny
The fight over AI Agents & Automation in therapy sessions arrives against a backdrop of repeated enforcement actions. In 2023, Kaiser settled with the California Department of Managed Health Care and paid a $50 million fine after regulators found violations in its behavioral health services, including timely access to care and provider oversight. Kaiser also agreed to invest $150 million to improve mental health services.
Then in February, Kaiser Foundation Health Plan reached a $31 million settlement with the U.S. Department of Labor over failures to provide timely and appropriate mental health services. In July, NUHW filed a complaint with state regulators alleging Kaiser's digital screening algorithm assesses patient answers and makes referrals without licensed clinician oversight, violating California law that requires clinicians to triage patients.
Kaiser did not attend the July supervisors hearing but submitted a letter stating it is "evaluating technology that can reduce administrative burden, improve care navigation, and help clinicians spend more time with patients." The letter added that patient safety, quality, privacy, and clinician oversight "remain paramount."
Where the labor fight stands
Clinicians held a one-day strike in March and have authorized a three-day strike, though no date has been set. The union says mediation in August did not produce a settlement, and no further sessions are scheduled. The San Francisco resolution carries no legal force but amplifies political pressure on Kaiser as the contract fight continues.
Why this matters for healthcare professionals
For clinicians and healthcare administrators, San Francisco's resolution signals growing institutional pushback against deploying AI in direct patient care without clear guardrails. The core dispute - whether algorithms can triage patients without licensed oversight - cuts to the center of how AI for Healthcare gets integrated into clinical workflows. If regulators uphold the union's complaint, it could set a precedent that mandates human clinicians remain in the loop for screening and referral decisions, even when AI tools handle the initial assessment. That has direct implications for staffing models, liability, and the pace of automation adoption across health systems.
Your membership also unlocks:
