Report: 1 in 5 American Teens Secretly Use AI Chatbots for Mental Health Help

More Than 9 in 10 Say AI Advice Helped, but Experts Urge Caution

An estimated 8 million American teenagers and young adults are turning to AI chatbots for emotional support, and most of them haven’t told a single person in their lives that it’s happening.

A nationally representative survey published in JAMA Pediatrics found that roughly 1 in 5 Americans aged 12 to 21 reported using AI chatbots for mental health advice as of 2025. That figure jumped sharply from just a year earlier, when a similar survey by the same research team found about 1 in 8 young people doing the same thing. That climb is unfolding against the backdrop of a well-documented youth mental health crisis in the U.S., where suicide is the second leading cause of death among those aged 5 to 24, and rates of depression and anxiety among teenagers are described by researchers as “alarmingly high.”

Hidden Use of AI Chatbots for Mental Health

Among young people who said they had used an AI chatbot for mental health advice, 63.3% had not mentioned it to anyone at all, whether a parent, a sibling, or a friend. Those who did open up were most likely to tell a friend (28.0%), while just 16.4% told a trusted adult such as a parent, teacher, or doctor.

That hidden pattern concerns the researchers behind the study. Without knowing a patient has been getting mental health guidance from an AI tool, doctors may miss chances to correct bad information, offer context, or connect young people to professional care. Authors of the study suggest that parents, clinicians, and educators start directly asking teenagers whether they’re using AI chatbots for emotional support, and have honest conversations about both the benefits and the real limits of those tools.

That concern has a concrete basis. While 91.7% of chatbot users in the survey rated the advice they received as somewhat or very helpful, the researchers flag an important caveat: AI systems are known to be agreeable and flattering by design, which means a teenager might walk away from a chatbot conversation feeling reassured regardless of whether the advice was actually sound. As the study authors wrote, “perceived helpfulness may reflect AI chatbots’ tendencies toward sycophancy and overflattery, rather than the quality of advice they provide.”

Inside a Nationally Representative Youth Survey

Conducted in November 2025, the survey included 1,009 young people between the ages of 12 and 21, drawn from RAND’s American Life Panel, a nationally representative survey group. After statistical adjustments to reflect the broader US population, the results represent an estimated 42.8 million American youth. Participants completed the survey online in English and received financial compensation for their time. Parents of minors provided consent for their children to participate, and the minors themselves agreed to take part.

Respondents were first given examples of AI chatbots, including ChatGPT, Google Gemini, Snapchat’s MyAI, Character.AI, and Meta AI, and asked whether they had ever used one. Those who said yes were then asked whether they had ever turned to a chatbot specifically for mental health advice, defined in plain terms as seeking help when feeling sad, angry, nervous, or stressed. Follow-up questions covered how often they used chatbots for this purpose, how helpful they found the responses, and whether they had told anyone about it.

Results varied in real ways across demographic groups. Females were about twice as likely as males to have used an AI chatbot for mental health advice. Older teens and young adults aged 18 to 21 were more than three and a half times more likely to have done so than the youngest group surveyed, those aged 12 to 14. Young people who had already spoken with a doctor about their mental health in the past six months were also more likely to have used a chatbot, suggesting that AI use may be layered on top of, rather than replacing, traditional care for some.

Among those who used chatbots for mental health advice at least monthly, Black youth were more likely to do so than White youth. Researchers offer two possible explanations: a sense that professionals are not as responsive to this group’s specific needs, or reduced access to professional mental health services. Because the survey wasn’t built to pin down a reason, both stay possibilities rather than findings, and the researchers note that the sample for this particular breakdown was small and warrants further study.

Regular Reliance on AI Chatbots for Mental Health

Of the young people who reported using AI chatbots for mental health advice, 42.8% did so at least monthly. Breaking that down: 26.3% used them at least once a month, 10.8% at least once a week, and 5.8% daily or almost daily. That level of steady engagement with a tool sitting outside any clinical oversight is part of why the researchers call this no marginal phenomenon, describing AI chatbots as now embedded in the mix of sources young people turn to for mental health information.

At 19.2%, chatbot use for mental health sits close in scale to the 19.8% of adolescents who reported receiving counseling from a mental health professional, a figure drawn from separate CDC data. Researchers are careful to point out these aren’t equivalent: professional counseling is formal clinical care, while AI chatbot use covers a much broader and looser range of interactions. Still, the comparison shows just how far this technology has entered the mental health conversation for young people.

A Youth Mental Health Crisis Meets the Chatbot

For all their limits, AI chatbots may be filling a genuine gap. Four in 10 teenagers with a major depressive episode in the past year report not receiving any mental health services, according to federal data cited in the study. Anxiety disorders have an even wider treatment gap. Against that backdrop, it’s not hard to understand why a teenager who feels overwhelmed at midnight might type their feelings into a chatbot rather than wait weeks for a therapy appointment.

But the very features that make AI chatbots appealing (always available, private, free of judgment, and no need to tell a parent) are also what make them risky when used without any adult awareness or professional oversight. Researchers are clear that AI chatbots built to offer therapeutic guidance are still in early stages, with few standards for measuring how well they perform and little transparency about how they’re made.

With 8 million young Americans already leaning on these tools, and most keeping it to themselves, the authors press an urgent case: parents and clinicians should start asking about chatbot use directly, rather than guiding care unaware of it.

Disclaimer: This article summarizes a cross-sectional survey, which measures behavior at a single point in time and captures what respondents say about themselves. A design like this can show how common a behavior is and which groups report it more often, but it cannot establish cause and effect, and it cannot confirm what actually happened in any individual chatbot conversation. The survey did not evaluate the quality or safety of the advice the chatbots provided. Figures reflect self-reported use among US youth aged 12 to 21 as of late 2025 and may shift as these tools and their user base change. Nothing here is medical advice. Anyone worried about a young person’s mental health should speak with a qualified clinician. This is a sensitive topic; readers who are struggling can reach free, confidential support in the US by calling or texting 988, the Suicide and Crisis Lifeline.

---