A recent study reveals that nearly a quarter of Ontario students use AI chatbots for emotional support, raising alarms among clinicians about potential impacts on mental health and the importance of addressing privacy and emotional development issues.
Children and adolescents are increasingly turning to AI chatbots for emotional support, a trend that is worrying clinicians because it may be a sign of deeper psychological distress. In a study of nearly 40,000 students in Ontario, published in JAMA Pediatrics on 31 August, seeking emotional support from AI emerged as a distinct marker of mental health problems, separate from feelings of loneliness or not mattering. The findings add to growing concern that chatbots are becoming part of how some young people cope with stress, sadness and isolation.
Peter Selby, a family physician and addiction specialist at the Centre for Addiction and Mental Health in Toronto, told Medscape News Canada that parents often notice only indirect signs, such as more screen time, changes in behaviour or a child seeming upset. He said a separate American study showed that young people often do not tell their doctors when they are using AI. That, he said, means clinicians should raise the issue during routine visits, including appointments for immunisations, and should ask about a child’s mental health before moving on to whether AI is part of their coping strategy.
The Ontario study found that about 21% of students reported using generative AI for emotional support. Girls, gender-diverse students and Black, Indigenous or other non-white students were more likely to do so than boys and White students, while older students and those using AI for school work were also more likely to report affective use. In unadjusted analysis, emotional support use was linked with higher emotional-problem scores and a near doubling in the prevalence of clinical emotional problems. The association remained after adjustment, while functional use, such as for homework, showed a smaller relationship with emotional difficulties.
The concern among specialists is not simply that young people are using a new tool, but that they may be replacing human relationships with something that is designed to be agreeable rather than challenging. Rachel Mitchell, a child and adolescent psychiatrist at Sunnybrook Health Sciences Centre in Toronto, told Medscape News Canada that doctors, teachers and parents should ask about chatbot use without judgement. She said the use becomes especially troubling when the bot starts standing in for a trusted adult or friend. Selby added that chatbots can create an echo chamber by simulating empathy while feeding back what the user wants to hear, which can be risky if the tool is treated as a form of therapy.
Rebecca Hicks, president of the Ontario Medical Association, told Medscape News Canada that many users do not realise the information they share may be stored and used to train commercial systems. She said some people may still choose to use chatbots after understanding that risk, but clinicians should make the privacy implications clear. For young patients whose use appears problematic, Selby said the conversation should start with the chatbot itself, so that diagnosis is not built on assumptions drawn from an AI-generated response. The goal, he said, is to work towards a diagnosis, not backwards from one.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
