Research involving nearly 40,000 Ontario students reveals rising use of AI chatbots for emotional support, with signs of increased emotional problems and concerns over substituting human care.
A new study in JAMA Pediatrics adds to growing concern about how children and teenagers are using artificial intelligence chatbots for emotional support. Researchers surveying nearly 40,000 students in Ontario, Canada, found that more than one in five said they used generative AI for advice or comfort about relationships, family or other personal matters, and that this use was associated with markedly higher emotional problem scores.
The study’s authors said the link was strongest among girls, gender-diverse students and Black, Indigenous and other racialised young people. Even after accounting for loneliness and a reduced sense of belonging, the association remained, suggesting the pattern was not explained only by those feelings. Separately, a JAMA Pediatrics study of about 1,000 adolescents and young adults in the United States found that roughly 19% had used AI chatbots for mental health advice in 2025, with many doing so regularly and most saying the responses were helpful, while more than half had not told anyone they were doing it.
Child psychiatrists warn that the appeal of an always-available chatbot may also hide risks. Consuelo Cagande, child and adolescent psychiatry division director at Northwell Zucker Hillside Hospital and Cohen Children’s Medical Center, told The New York Post that the trend was deeply worrying because young people may be replacing human support with software that has no clinical safeguards, follow-up or crisis response. She said that if children do use chatbots, parents should supervise closely and schools should teach digital literacy so young people understand the limits of AI-generated advice.
The findings arrive as regulators and health experts debate how to protect minors from the harms of AI while still allowing responsible use. Other recent coverage has noted that some teenagers turn to chatbots not just for advice but for entertainment, companionship and even romance, raising fresh questions about emotional dependence, misinformation and blurred boundaries. For parents and caregivers, the message from the emerging research is blunt: AI may be a tool, but it should not be the first place a distressed child turns for help.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
