New research published in JAMA Pediatrics has revealed a significant association between the use of artificial intelligence chatbots for emotional support and an increase in emotional problems among adolescents. The study surveyed nearly 40,000 students in grades four through 12 in Ontario, Canada, finding that over 21% reported using generative AI for "affective purposes," such as seeking emotional guidance or advice regarding personal relationships.
Students who utilized chatbots for these reasons exhibited "significantly higher emotional problem scores," with nearly double the likelihood of experiencing clinically significant symptoms of distress. The study also identified disparities in these effects, noting that the link between AI use and emotional issues was more pronounced among girls, gender-diverse students, and Black, Indigenous, and other racialized students.
Experts express concern over this trend. Dr. Consuelo Cagande, division director for child and adolescent psychiatry at Northwell Zucker Hillside Hospital and Cohen Children’s Medical Center, described young people turning to AI chatbots for emotional help as "very concerning." She observed a progressive substitution of traditional social support systems and treatment models with AI algorithms.
Cagande highlighted the "lack of safeguards or clinical follow-ups or crisis safeguards for those at risk who are using chatbots." The research indicated that young individuals seeking emotional support from AI also reported feelings of loneliness and a lack of perceived importance, although the link between AI use and emotional problems persisted even when these factors were statistically controlled.
Concerns exist that some adolescents may fear judgment from humans when discussing their problems, leading them to confide in AI. Cagande recommends supervision for any young person using AI chatbots, alongside education on digital literacy and AI usage due to the potential for misinformation. She stressed that generative AI should not be the primary resource for emotional distress or mental health concerns, but rather a supplemental tool.
If AI chatbots are used, Cagande advises that any information received about mental health issues should be shared with a clinician rather than accepted as final advice. Traditional support systems, including consultations with physicians for resource referrals and mental health professionals, as well as guidance from teachers and school counselors, should remain the primary avenues for support.
While the European Union has implemented bans on certain health-threatening AI technologies, the United States lags in comprehensive federal regulations governing artificial intelligence. This new research underscores the potential mental health implications of AI technologies on vulnerable youth populations.