發生了什麼事
California Governor Gavin Newsom has vetoed Senate Bill 903, a legislative measure that aimed to regulate the use of artificial intelligence within the field of therapy. The bill, which had passed the state Senate unanimously and faced only one dissenting vote in the Assembly, would have prohibited therapists from using AI tools to make clinical decisions or evaluate a patient's emotional state without direct oversight from a licensed professional. Additionally, the legislation sought to restrict the advertising of therapy services provided by chatbots.
Governor Gavin Newsom vetoed Senate Bill 903 on September 30, 2026, the final day for action on the year's legislative package. The bill was authored by State Senator Steve Padilla.
The legislation would have mandated that any AI tool used to evaluate a person's emotional state or make clinical decisions in a therapy context must be reviewed and approved by a licensed professional.
The bill also included provisions to regulate how therapy services provided by chatbots are advertised to the public.
In his veto message, Newsom stated that the bill was 'overly broad' and would 'drastically limit a clinician’s use of tools that benefit the delivery of care today, including by requiring routine screening determinations to receive direct approval.'
為什麼這很重要
The veto highlights the ongoing tension between state-level efforts to implement and the practical concerns of healthcare providers regarding the integration of new technology. While proponents, including the California Association of Marriage and Family Therapists and the National Union of Healthcare Workers, argued the bill was necessary to ensure patient safety and professional accountability, the governor and industry opponents argued that the language was too restrictive. The veto prevents a significant regulatory hurdle for the adoption of AI-assisted clinical tools, leaving the current landscape of AI in mental health largely governed by existing professional standards rather than specific new statutory mandates.
The veto represents a significant setback for advocates who sought to establish clear legal boundaries for AI in sensitive mental health settings. Supporters, such as the National Union of Healthcare Workers, emphasized the need for assurance that therapy services are provided by qualified humans.
Conversely, the California Hospital Association and other trade groups opposed the bill, arguing that its requirements would create unnecessary administrative burdens and impede the use of AI tools that can assist therapists in their daily practice.
The decision underscores the difficulty of crafting AI regulation that balances safety and consumer protection with the potential for technological innovation in healthcare delivery.
This action follows other recent legislative efforts in California to regulate AI, including the signing of laws regarding AI in the workplace and for children, suggesting that the state remains a primary testing ground for AI policy.
互動機制：它實際上是如何運作的
以互動方式探索這項發展背後的基礎技術。
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
接下來看什麼
The veto leaves a gap in specific state-level regulation for AI in mental health, which may lead to future legislative attempts to address the issue with more targeted language. Observers should monitor whether the California Legislature introduces revised versions of the bill in the next session that address the governor's concerns about the 'overly broad' nature of the original text. Additionally, the impact of this veto on the broader landscape of AI-driven healthcare services in California remains a key area for ongoing scrutiny.
Future legislative sessions may see new attempts to regulate AI in therapy, potentially with more specific language designed to satisfy the governor's concerns regarding the scope of the original bill.
The absence of this law means that the use of AI in mental health will continue to be governed by existing professional licensing boards and general consumer protection laws, rather than specific AI-focused statutes.
Stakeholders will likely continue to debate the appropriate level of human oversight required for AI-assisted mental health tools as these technologies continue to evolve and proliferate.