加州立法者正推进SB 903法案，力图对人工智能在心理健康领域的应用设置监管边界[1]。该法案禁止企业将聊天机器人宣传为治疗工具，并要求AI系统在做出治疗决定前必须接受执业专业人士的审查[1]。此外，法案还规定医疗提供者在使用AI工具录制治疗会议或进行分诊前，需要首先获得患者的明确许可[1]。
这一立法举措针对的是越来越多人转向AI寻求心理咨询帮助的现象。据报道，约八分之一的青少年和年轻人使用聊天机器人获取心理健康建议[1]。OpenAI曾报告称，每周约有120万用户在ChatGPT上分享自杀念头[1]。加州参议员Steve Padilla表示，"AI可用于行政目的和支持临床医生，但不能进行心理治疗"[1]。此前，伊利诺伊州已成为首个通过禁止AI用于治疗服务法律的州[1]。
California legislators are advancing Senate Bill 903, which seeks to restrict how artificial intelligence chatbots can be used in mental health care [1]. The legislation prohibits companies from marketing chatbots as therapeutic tools and requires that any therapeutic decisions made by AI be reviewed by licensed professionals before implementation [1]. The measure also mandates that healthcare providers obtain patient consent before using AI tools to record therapy sessions or conduct patient triage [1].
The push for regulation comes as growing numbers of young people turn to AI for mental health support. According to reports, approximately one-eighth of teenagers and young adults use chatbots to seek mental health assistance [1]. OpenAI has disclosed that roughly 1.2 million users share suicidal thoughts on ChatGPT each week [1]. Senator Steve Padilla, a key advocate for the bill, stated that "AI can be used for administrative purposes and to support clinicians, but it cannot conduct psychotherapy" [1]. Illinois became the first state to pass legislation prohibiting AI use for therapeutic services last year [1].