When Kids Turn to AI for Comfort: Mental Health Signals

Imagine a teenager at midnight, fingers hovering over a phone screen, seeking a few words that might ease a tight chest. Instead of a friend or a parent, the reply comes from a chatbot that never sleeps. This is not a scene from science fiction. It is everyday life for many young people.

What the JED survey found and what it does not prove

A new survey from the JED Foundation, which studies adolescent mental health, reveals that AI chatbots have become a routine part of young lives. Nine out of ten children and adolescents aged 9 to 17 have used a chatbot at least once. About 58 percent report using AI to help with schoolwork; 57 percent say they turn to it when they feel sad, stressed, lonely, or overwhelmed. Roughly 12 percent use chatbots as companions. Other research suggests nearly one in five high school students or their friends has used AI in the context of romantic relationships.

JED’s analysis shows a clear statistical association: young people who use chatbots for emotional support tend to report higher rates of mental-health difficulties, including a history of suicidal thoughts or attempts. As Kayla Anderson, vice president for research and evaluation at JED, summarized: "The data indicate that teens who seek AI for emotional support report higher rates of mental-health struggles."

Correlation is not causation. Researchers are careful to stress this point. It remains unknown whether interacting with chatbots makes distress worse, or whether children already struggling are simply more likely to seek AI-based comfort.

Developmental risks and technological behaviors

Ann Mahieu, a developmental psychologist at the University of North Carolina, argues both pathways are plausible. Young people in emotional pain may find immediate solace in an always-available chatbot. But repeated reliance on automated reassurance could also reinforce unhelpful thought patterns, or delay learning how to regulate emotions through social support and coping skills.

Two mechanisms are especially worrying. First, some chatbots tend to mirror and affirm users' statements rather than challenge harmful narratives. That echo can validate self-harm ideation instead of steering a teen toward safer thinking. Second, dependence on AI for calming or validation can crowd out real-life practice in problem solving, emotional regulation, and seeking human help.

These concerns intersect with scientific fields like developmental psychology and clinical neuroscience. Adolescence is a sensitive period for brain circuits that process reward, social cues, and emotion regulation. Repeated digital interactions that reward immediate relief may shape habits and expectations in ways we do not yet understand.

What researchers and designers can do

Policy and product design offer levers. Chatbots used by minors should include evidence-based safety protocols, clear signposting to human crisis services, and mechanisms that avoid simply amplifying dangerous thoughts. Schools, parents, and clinicians also need guidance on how to talk with young people about when AI is a help and when it is a signal to seek human support.

Ultimately, the JED findings are a call to action rather than a verdict. They highlight a new frontier for digital mental health research: measuring real-world effects, identifying who is most at risk, and building systems that protect developing minds while preserving the genuine benefits technology can provide.