California Moves to Ban AI Chatbots from Advertising as Therapists

California lawmakers are advancing Senate Bill 903 to prevent AI chatbots from being marketed as therapists and to strictly limit their role in clinical mental health settings. The legislation ensures that no automated tool can claim to provide professional psychotherapy without a licensed human clinician remaining accountable for the patient's care.

Under the proposed law, companies are prohibited from advertising AI chatbots as providing "therapy" or "psychotherapy" unless a licensed professional is responsible for the care. The bill forbids AI from independently interacting with clients, detecting emotions, or generating treatment plans without the approval of a licensed professional. For providers who use AI for recording, transcribing, or triaging care, the bill mandates written disclosure and affirmative consent from the patient. Enforcement would be handled by the Department of Consumer Affairs, with potential civil penalties reaching $10,000 per violation.

The Crisis of Access and Safety

The push for regulation comes as Californians face a severe shortage of mental health providers, with nearly one-third of the population living in areas where provider-to-patient ratios are insufficient. This gap has driven a surge in the use of AI chatbots for emotional support. While some research suggests these tools can reduce symptoms of anxiety and depression, safety advocates warn of significant risks. Experts have cautioned that chatbots often fail to recognize crisis warning signs, posing a particular danger to teenagers.

A Shift in Regulatory Strategy

This legislation marks a strategic shift in how the state handles generative AI, moving away from outright bans toward the regulation of clinical claims. By targeting the specific terminology of "therapy," California is attempting to protect vulnerable users from unlicensed AI services while still permitting the existence of general-purpose AI tools. The distinction between a licensed clinician and an automated response can be life-altering for patients in crisis.

Legal Precedents and Next Steps

The bill arrives as AI developers face increasing legal scrutiny; wrongful death lawsuits have already been filed in California federal courts alleging that chatbot makers contributed to user suicides. Having passed the Senate unanimously in May, SB 903 is currently on the Assembly Appropriations Committee's suspense file. If signed into law, it will set a significant precedent for how other states manage the intersection of regulated healthcare professions and autonomous technology.