Nearly one in five American teenagers are quietly reaching for AI chatbots when they feel sad, angry, nervous, or stressed—and most aren't telling anyone about it. A new study published in JAMA Pediatrics reveals that as the youth mental health crisis deepens across the United States, an unprecedented wave of young people are turning to generative AI models like ChatGPT and Google Gemini for emotional support, often in complete secrecy from parents, friends, and trusted adults.
The timing is striking. About one in three U.S. high school students report poor mental health most or all of the time, according to recent CDC data. One in five have seriously considered suicide, and nearly one in ten have attempted it. In this landscape of crisis, AI chatbots have become a surprising refuge—and a hidden one. Researchers surveyed 1,009 young people aged 12 to 21, with results weighted to represent over 42 million youth nationwide, and found that approximately 20% have used an AI chatbot for mental health advice at least once.
What's most striking is not the uptake, but the silence surrounding it. Among those using chatbots for support, 63.3% kept the practice entirely hidden from others. This secrecy reflects something troubling: young people are increasingly managing their mental health in private conversations with machines, away from the oversight and guidance of the adults in their lives. More than 40% of users turn to these chatbots at least once a month, and nearly 6% use them daily or almost daily.
The growth is rapid. Just a year ago, only 13% of young people reported using AI chatbots for mental health support. That number has nearly doubled to 20%, signaling an accelerating trend that researchers say deserves urgent attention from parents and clinicians. The study found that older teens and girls are more likely to use these tools, and usage is particularly high among those who have already spoken to a physician about their mental health within the previous six months.
Nearly 92% of users reported finding the advice helpful—a figure that warrants caution, researchers warn. The sense of helpfulness may stem not from the quality or accuracy of the advice itself, but from chatbots' tendency to be reflexively agreeable and flattering, characteristics that can mask poor guidance or misinformation. A chatbot that always validates a user's feelings might feel supportive in the moment while actually reinforcing harmful thought patterns.
Demographic trends reveal interesting patterns. Young people most likely to turn to AI for mental health support are those already engaged with the healthcare system, suggesting the tools aren't simply replacing professional help—at least not yet. Instead, they're filling gaps, becoming an unsupervised companion in the mental health information ecosystem that young people navigate daily.
As AI chatbots embed themselves deeper into how young people manage their mental health, researchers emphasize that parents and clinicians need to shift from resistance to conversation. Rather than forbidding chatbot use, adults need to openly discuss these tools, help set realistic expectations about their limitations, and ensure young people know when and how to connect with actual mental health professionals. The secrecy surrounding chatbot use suggests that many young people feel they cannot talk about their struggles with the people closest to them—a gap that technology is filling, but not solving.