A nationally representative study published in JAMA Pediatrics has found that roughly 1 in 5 American adolescents and young adults between the ages of 12 and 21 are turning to AI chatbots for mental health support, and nearly two-thirds of them have not told a single adult, parent, or friend that they are doing so.
The findings, drawn from a survey weighted to represent approximately 42 million young people across the United States, signal a rapid and largely invisible shift in how the current generation manages emotional distress.
The study’s lead author, Ryan K. McBain, Ph.D., M.P.H., of RAND, put the stakes plainly in a statement reported by Ophthalmology Advisor. “AI chatbots are already part of how many young people seek advice about their mental health,” McBain said.
“The speed of growth is attention-grabbing, but so is the fact that most young people who use these tools for mental health advice say they are not telling anyone.”
How Big Is This Trend, And How Fast Is It Growing?
The numbers are striking on their own, but the trajectory makes them more urgent. A comparable RAND survey conducted just one year earlier put the rate of AI chatbot use for mental health purposes at roughly 13%, meaning the figure has climbed to nearly 20% in a single year.
Translated into real population terms, an estimated 8.2 million young people are now seeking emotional guidance from AI tools, a number that now nearly equals the share of young people receiving care from a licensed mental health professional.
Among those who do use AI chatbots for mental health advice, the habit is not casual. About 43% reported seeking that guidance at least once a month, and 5.8% reported daily or near-daily use, according to the study. The platforms teens are turning to include widely available tools such as ChatGPT, Gemini, Character.AI, and Meta AI.
The research team, whose institutional affiliations span RAND, Mass General Brigham, Harvard Medical School, MIT Media Lab, and Brown University School of Public Health, used statistical weighting on a base sample of 1,009 participants to ensure the results reflected the broader national population of young people ages 12 to 21.
The Secrecy Problem, And Why It Matters
Perhaps the most alarming finding is not the usage rate itself, but the silence surrounding it. Roughly 63% of young people who used AI chatbots for mental health guidance reported that they had not disclosed this to anyone, including parents, friends, or healthcare providers.
That secrecy removes any safety net that might otherwise catch a teen who receives harmful, inaccurate, or clinically inappropriate advice.
The overwhelming majority of users, about 91 to 92%, rated the AI advice they received as somewhat or very helpful. But researchers caution that satisfaction figures like these may say more about how chatbots are engineered to respond than about whether the underlying guidance is sound or clinically appropriate.
Generative AI systems are built to produce supportive, engaging responses, which can make users feel heard and validated without guaranteeing that the information is accurate, evidence-based, or appropriate for serious mental health situations.
As HealthDay noted in its coverage of the study, researchers flagged that high helpfulness ratings may say more about how chatbots are designed than about whether the advice is genuinely sound.
The Mental Health Crisis Driving Teens To Their Screens
To understand why teens are turning to AI in such numbers, it helps to look at the mental health landscape they are navigating. CDC data cited in the research context paints a sobering picture: close to one-third of high school students said they had experienced poor mental health for most or all of the previous month.
Suicide-related statistics are equally troubling, with more than one in five reporting that they had seriously considered ending their life and roughly one in ten saying they had made at least one attempt. Against that backdrop, the appeal of an always-available, judgment-free digital listener becomes easier to understand, even if the risks remain real.
Dr. Riana Elyse Anderson, a licensed clinical and community psychologist and associate professor at Columbia University’s School of Social Work, observed this dynamic firsthand while spending time with teenagers at friends’ homes. Writing for HuffPost, she described watching teens turn to AI chatbots the moment emotional discomfort surfaced. She noted that the teens she observed did not want to schedule an appointment for later or wait to address their distress; they wanted something that could give them support almost instantly.
That desire for immediate relief is not simply a character flaw in today’s teens. It reflects a genuine mismatch between the speed at which young people experience distress and the pace at which the traditional mental health care system can respond.
Appointment wait times, insurance barriers, and the stigma of seeking help all push teens toward whatever is fastest and most private.
Real Risks Behind The Helpful Ratings
The secrecy surrounding AI mental health use carries consequences that go beyond missed clinical opportunities.
Anderson, writing in her HuffPost piece, raised a concern that many clinicians share: “Young people are also less likely to discuss the advice from [ChatGPT] with loved ones, leading to potentially deadly consequences, such as in cases where teen suicides have been linked to AI chatbots.”
She also highlighted a dimension of the problem that broader coverage often overlooks: the potential for AI tools to produce racially biased guidance.
For Black youth already navigating elevated rates of online racism and higher rates of suicidal ideation relative to their peers, AI chatbot advice drawn from biased training data could deepen existing mental health disparities rather than close them.
Researchers also note what they call a “compassion illusion,” in which AI-generated responses create a convincing impression of empathy and understanding despite the fact that the system has no clinical judgment, no duty of care, and no ability to escalate a crisis appropriately.
What Parents And Clinicians Can Do Right Now
The goal should not be to ban or shame teens out of using AI tools, but to bring those conversations into the open. Parents and clinicians who ask direct, open-ended questions about how teens are using AI, including for emotional support, are more likely to learn what is actually happening and to offer meaningful guidance alongside it.
Experts stress that the conversation itself, not the technology, is where parents can make the most difference.
Anderson offered a practical framework for parents in her HuffPost piece: model the behavior you want to see, demonstrate how to work through a problem out loud, and ask questions that invite real conversation rather than yes-or-no answers.
Asking a teen what they have been talking to an AI chatbot about recently is not an accusation; it is an opening.
The study’s authors also recommend that healthcare providers routinely ask adolescent patients whether they are using AI tools for mental health purposes, just as they might ask about social media use or sleep habits.
Given that teens who had already spoken with a physician about mental health were more likely to also use AI chatbots, the two behaviors are not mutually exclusive, and clinical conversations can help shape how teens use these tools safely.
There Are Resources For Parents
For parents who want to understand the broader landscape of AI tools teens are engaging with, including the risks that extend beyond mental health, our coverage of AI-powered products marketed to young people offers important context.
And if your teen is already showing signs of relying on digital tools for emotional support, our piece on why AI cannot replace a real therapist is worth reading together.
What this study really reveals is a gap, not just between teens and adults, but between the speed of technological adoption and the pace of the systems meant to keep young people safe.
AI chatbots did not create the youth mental health crisis, but they are now deeply embedded in how millions of teens are trying to cope with it.
The question for parents is not whether to fight that reality, but how to stay present within it, curious rather than alarmed, and close enough to the conversation that your teen knows a real human is also available, any time, no appointment needed.
