AI Chatbots Surge as a Primary Source of Mental-Health Advice for U.S. Adolescents

A nationally representative survey of 1,009 Americans aged 12-21, weighted to 43 million youths, found that 19 % (? 8 million) have turned to AI chatbots such as ChatGPT, Gemini, Character.AI or Meta AI for help when feeling sad, angry, nervous or stressed. Among users, 40 % report at least monthly use, and 91 % rate the advice “somewhat or very helpful.” Nearly two-thirds (63 %) have not disclosed this use to any adult. Usage is higher among females, 18-21-year-olds, and those who have spoken with a physician in the prior six months.

The United States faces a persistent youth mental-health crisis: roughly 40 % of teens report depressive symptoms and 10 % have attempted suicide. Only about half of those who need care receive it, leaving a service gap that AI tools are filling despite lacking clinical validation.

• 19 % (19.2 % per a separate report) of adolescents and young adults use AI chatbots for emotional support.

CNN

• ? 8 million youths nationwide.

CNN

• 40 %–43 % use chatbots at least once a month.

• 91 %–92 % consider the advice helpful.

CNN

• 63 % do not tell parents, clinicians or teachers.

CNN

• Higher prevalence among females and ages 18-21.

AI chatbots offer 24/7, non-judgmental access, which can lower barriers for reluctant users. However, they can hallucinate, provide inaccurate or unsafe advice, reinforce harmful beliefs, and delay professional intervention—especially for severe depression, anxiety, self-harm or suicidal ideation. The perceived helpfulness may reflect chatbot flattery rather than therapeutic efficacy.

CNN

• Dr. Leana Wen (GWU associate professor) notes the tools are “available at any hour… respond instantly… do not appear judgmental and can feel private,” but stresses “accuracy… hallucinations… inability to recognize urgent crises.”

CNN

• Kathleen Young, PhD, Director of Behavioral Health at Novant Health, says AI can “supplement… stress-management techniques” but “is not the same as a therapist.”

Study finds nearly 1 in 5 teens use AI chatbots for mental health support

• C. Vaile Wright, senior director of the APA’s Office of Healthcare Innovation, asserts “there is absolutely no consensus… that AI chatbots can serve as a replacement for therapy.”

• Jonathan H. Cantor, RAND senior policy researcher, warns that “many young people appear to be using AI chatbots for mental-health advice privately, without the knowledge of parents, clinicians or other adults.”

• Legislative response: the Senate Judiciary Committee advanced the GUARD Act (age verification, ban on AI companions for minors); Utah enacted HB 452 regulating AI-based mental-health tools.

Teens are Relying on AI for Emotional Support. Policymakers Must Respond

Experts cite three core risks: (1) factual inaccuracy and hallucination; (2) reinforcement of distorted thinking through sycophantic responses; (3) potential for users to postpone or avoid professional care. A Center for Countering Digital Hate test found 47 % of AI replies to fabricated 13-year-old profiles suggested harmful behaviors, including substance use and suicide.

The Dangers of AI Chatbots for Teen Mental Health

Teens describe chatbots as “private, immediate, and non-judgmental,” making them easier to approach than parents or counselors. Because most do not disclose use, adults often remain unaware of the influence on youths’ coping strategies.

The Dangers of AI Chatbots for Teen Mental Health

• Helpful-rating figures vary (91 % vs 92 %).

CNN

• Overall usage reported as 19 % vs 19.2 %.

CNN

• Monthly-use rates cited as 40 % vs 43 %.

• No longitudinal data demonstrate symptom improvement or safety outcomes, leaving a critical evidence gap.

CNN

• “These tools are available at any hour of the day, they respond instantly, they do not appear judgmental and they can feel private.” — Dr. Leana Wen, emergency physician

CNN

• “These systems can hallucinate and provide incorrect information.” — Dr. Leana Wen

CNN

• “You certainly can put in symptoms of the way that you’re feeling and ChatGPT would give you a diagnosis, but it doesn’t necessarily have the skills to be able to understand what follow up questions would be necessary to really understand what’s going on, as well as the context that the person is in,” — Kathleen Young, PhD

Expert urges caution as young people seek mental health support from AI

• “Many young people appear to be using AI chatbots for mental health advice privately, without the knowledge of parents, clinicians or other adults,” — Jonathan H. Cantor, RAND

• “There is absolutely no consensus in the field that AI chatbots can serve in any way as a replacement for therapy,” — C. Vaile Wright, APA

• “AI companions socially distance us from real people and can discourage real-world connections,” — Don Grant, PhD

The Dangers of AI Chatbots for Teen Mental Health

Policymakers are moving toward stricter regulation (GUARD Act, Utah HB 452) while researchers call for validated, evidence-based AI interventions. Clinicians and parents are urged to discuss AI use openly, verify information, and treat chatbots as supplemental tools rather than substitutes for professional mental-health care.

CNN