AI chatbots linked to higher emotional distress among teens, study finds
A new study of nearly 40,000 Ontario students shows that using generative AI for emotional support is associated with significantly higher emotional problem scores, especially among girls, gender-diverse and racialized youth.
In a JAMA Pediatrics paper, investigators examined the AI habits of almost 40,000 Ontario pupils in grades four through twelve. Over 21% reported turning to generative AI for emotional advice or support, a practice that correlated with markedly higher emotional problem scores and nearly double the odds of clinically significant distress. The link was most pronounced among girls, gender-diverse learners, and Black, Indigenous, and other racialized students.
Dr. Consuelo Cagande, a child and adolescent psychiatry division director, expressed serious concern about youths substituting traditional social networks and treatment models with chatbot algorithms that lack crisis safeguards. She urged that AI be used only as a supplemental tool, with clinicians informed of any chatbot-derived advice, and emphasized the need for supervision, digital-literacy training, and reliable mental-health referrals as first-line support.
Why it matters
The findings highlight mental-health risks of AI chatbots for vulnerable youth, urging safer digital practices and professional oversight.
In this story
