California's Senate Bill 903, which would prohibit AI chatbots from providing psychotherapy without human oversight, has cleared the Assembly Appropriations Committee. The bill aims to prevent AI from making independent therapeutic decisions, despite concerns that it could delay care amid worker shortages.
The bill reflects a growing tension between the potential benefits of AI in healthcare and the need for human oversight and ethical considerations, particularly in sensitive areas like mental health. Its passage could set a precedent for other states and influence how AI is integrated into therapeutic practices.
California is moving closer to prohibiting AI chatbots from acting as therapists and making mental health decisions without human supervision. Senate Bill 903, introduced in January, cleared the Assembly Appropriations Committee with a unanimous 13-0 vote on Thursday and has been ordered to a third reading. The bill, also known as the Wellness and Oversight for Psychological Resources Act, would restrict AI in psychotherapy to support roles, prevent chatbots from impersonating therapists, and mandate clinician oversight and patient consent. It also includes protections for therapy data and exemptions for peer support, religious counseling, research, and approved medical tools.
State Senator Steve Padilla, the bill's author, stated that AI algorithms are not equipped to replicate the skills and training of human therapists and emphasized the need for guardrails to prevent potential harm to patients. The bill previously passed the Senate with a 39-0 vote in May and cleared two Assembly policy committees with strong bipartisan support.
However, critics like Robert Boykin of TechNet argue that the requirement for clinician involvement could create delays in care, particularly in counties facing shortages of behavioral healthcare workers. Boykin suggested that SB 903 could place a 'clinician bottleneck' in front of intake and screening tools that help patients access care more quickly.
This legislative development occurs amid growing concerns about AI's impact on mental health. A June survey by the American Psychological Association revealed that 77% of psychologists have patients who use AI for mental health purposes. Furthermore, research from King's College London, the Protestant University of Applied Sciences, and the City University of New York indicated that chatbots can reinforce delusions through excessive agreement and personalized responses, likening the effect to an 'echo chamber of one' that lacks the corrective influence of real-life social interactions.