AI and Mental Health among Youth (Senate Bill 903)
Sebastian Cañas
One in every five teens use AI chatbot therapy that has been known to fail to recognize suicidal intent, offer dangerous suggestions, validate harmful thoughts, approve avoidant behaviors instead of challanging teens to heal, and give inappropriate advice. These factors can lead to teens harming themselves and potentially committing suicide. Additionally, chatbot developers can see the conversations between the victim and the chatbot.
Imagine you are stressed about schoolwork. You open an AI chatbot and explain what you are stressed about. The chatbot instantly replies with what may sound like perfect advice, but it may be only telling you what you want to hear. It may feel like a real conversation, yet there is nobody on the other side of the screen–just an algorithm processing code.
To protect teens, California lawmakers introduced Senate Bill (SB) 903, or the Wellness and Oversight for Psychological Resources Act. This legislation sets boundaries to make sure that AI chatbots cannot replace real, human mental health professionals.
The new rules of the legislation include:
Banning AI chatbot therapy: Companies cannot market independent AI chatbots directed as therapy
Any service considered therapy must be led by a living, licensed therapist.
Human therapists cannot use AI to record sessions without the parent’s consent.
Tech tools can only handle administrative work such as scheduling or typing up office notes.
Here at the CAYC, we focus heavily on youth well-being. While automated apps seem like a quick fix, I argue that complex emotional struggles require real human empathy, not automated code that just tells you what you want to hear. Under SB 903, California enhances safety regulations that help keep teens from using AI chatbots as therapy.
SB 903 will keep the tech companies developing AI chatbot therapy in check, with up to a 10,000 dollar fine for tech companies who break any of the new rules for SB 903. But, even though this may stop the companies from adding specified AI chatbot therapy, people still can type their feelings into a regular AI, such as copilot or Chat GPT, and get answers back. Further steps must be taken to prevent this from happening such as a pop up warning when someone asks AI questions that suggest the need for mental help.
Image credit: https://medicalxpress.com/news/2026-06-teens-ai-chatbots-mental-health.html