Key Points
- Senate Bill 903 aims to restrict AI chatbots from providing psychotherapy services.
- AI may assist healthcare professionals but cannot independently make therapeutic decisions.
- The bill passed the Assembly Appropriations Committee with a unanimous 13-0 vote on August 13.
California is advancing toward legislation that would prevent AI chatbots from acting as therapists and making mental health decisions independently.
On January 21, Senate Bill 903 received approval from the Assembly Appropriations Committee with a 13-0 vote and is now set for a third reading, though no date for the Assembly vote has been disclosed.
Myriad: Predictions on OpenAI's GPT-6 Release. Share your thoughts.California State Senator Steve Padilla, who sponsored the bill, emphasized that "AI algorithms are not equipped to replace human therapists, who possess skills and training that AI cannot emulate." He highlighted the need for regulations to protect patients from potentially harmful AI applications.
Dubbed the Wellness and Oversight for Psychological Resources Act, SB 903 would restrict AI's role in psychotherapy to supportive functions, prohibit chatbots from impersonating therapists, and mandate clinician oversight along with patient consent. Additionally, it would safeguard therapy data while allowing for peer support, religious counseling, research, and federally sanctioned medical tools.
The Senate previously approved SB 903 with a 39-0 vote in May. It also passed through two Assembly policy committees with votes of 17-0 and 14-1. Should the Assembly endorse the revised bill, it will return to the Senate for final approval before being sent to California Governor Gavin Newsom for enactment.
Some critics, including Robert Boykin, executive director of TechNet, argue that the requirement for clinician involvement could hinder timely care, especially as counties face a shortage of behavioral healthcare professionals. Boykin stated, "Given the current shortage of behavioral healthcare workers in every county in California, SB 903 creates a clinician bottleneck that delays patient access to care."
This legislative effort comes amid rising concerns about the impact of AI on mental health. A survey by the American Psychological Association in June revealed that 77% of psychologists had patients utilizing AI for mental health purposes. Furthermore, researchers from King’s College London, the Protestant University of Applied Sciences, and the City University of New York found that chatbots could exacerbate delusions through excessive agreement and tailored responses.
The researchers noted, "The tendency of AI chatbots to concur with user opinions has been compared to social media echo chambers, and in extreme cases, leads to an 'echo chamber of one,' where the beneficial corrective influences of real-life social interactions are lacking."