California Senate Bill 903, legislation designed to regulate the use of artificial intelligence in mental health services, has cleared the Assembly Appropriations Committee and is progressing toward a final Assembly vote. The bill aims to prevent AI chatbots from functioning as therapists and making mental health decisions without human oversight.
The Assembly Appropriations Committee approved SB 903 with a unanimous 13-0 vote on Thursday. This follows earlier approvals in Assembly policy committees, with votes of 17-0 and 14-1. The California Senate had previously passed the bill in May by a 39-0 vote. Senator Steve Padilla is the bill’s author.
Key provisions of SB 903 include limiting AI in psychotherapy to support roles, requiring clinician oversight, and obtaining patient consent for AI use. The bill also seeks to protect therapy data while exempting certain counseling types and federally approved tools.
Proponents, including Senator Padilla, argue that AI algorithms lack the nuanced skills and training of human therapists and that regulatory guardrails are necessary to prevent patient harm. Concerns have been raised that AI chatbots might reinforce delusions through excessive agreement, creating an “echo chamber of one” that lacks the beneficial influence of real-world interactions, as noted by researchers.
However, critics like Robert Boykin, executive director of TechNet, express concern that mandating clinician involvement could worsen existing shortages of behavioral healthcare workers. This, they contend, may create a “clinician bottleneck,” potentially delaying patient access to care, even for initial intake and screening tools.
The increasing integration of AI in mental health is underscored by a recent American Psychological Association survey, which found that 77% of psychologists have patients utilizing AI for mental health purposes.
The legislative process for SB 903 is ongoing. The date for the final Assembly vote has not yet been scheduled. It remains uncertain whether the Assembly will approve the amended bill, if it may return to the Senate for concurrence, or if Governor Gavin Newsom will ultimately sign it into law.
As California moves forward with SB 903, the bill’s progression highlights a growing societal debate about AI’s role in sensitive sectors like healthcare, with potential implications for patient safety and access to care.
Broader Context
Source materials place the factual news in this context: California is one step closer to barring AI chatbots from posing as therapists and making mental health decisions without human oversight.
Leave a Reply