A major new study of nearly 40,000 students is raising concerns about the growing tendency of children and teenagers to use artificial-intelligence chatbots as substitutes for human emotional support. Researchers found that 21.1% of surveyed students had used generative AI for emotional help or relationship advice, and those students were nearly twice as likely to exhibit clinically significant emotional problems as peers who did not use AI for those purposes. The research does not prove that chatbots cause psychological distress—emotionally troubled young people may be more inclined to seek AI assistance in the first place—but the relationship remained significant even after accounting for loneliness and feelings of insignificance. The findings reinforce a broader concern: AI may be useful for information and education, but allowing algorithms to become confidants for vulnerable children risks weakening the role of parents, families, counselors, teachers, and other human relationships precisely when young people need them most.
Key Takeaways
- More than one in five students surveyed—8,408 of 39,761—reported using generative AI for emotional support or relationship advice, demonstrating that chatbot use has already expanded well beyond schoolwork and information gathering.
- Students using AI for emotional purposes were nearly twice as likely to meet the threshold for clinically significant emotional problems, with rates of 57.7% among affective AI users compared with 29.2% among nonusers.
- The study establishes an association rather than causation, but independent research showing substantial youth reliance on chatbots for mental-health advice strengthens the case for parents and institutions to distinguish between AI as a practical tool and AI as a substitute for genuine human emotional relationships.
In-Depth
A survey of 39,761 students in grades 4 through 12 is adding weight to concerns that artificial intelligence is moving beyond homework assistance and into a sensitive role: emotional refuge. Researchers found that 21.1% of students reported using generative AI for emotional support or relationship advice. Those students showed substantially higher emotional-problem scores and were nearly twice as likely to meet the study’s threshold for clinically significant emotional problems: 57.7% compared with 29.2% among students who did not use AI this way.
The findings do not establish that chatbots caused the distress. Because the research was cross-sectional, troubled young people may simply be more likely to seek anonymous, always-available assistance from AI. That distinction matters. Still, the association remained significant after researchers accounted for loneliness and feelings of mattering, suggesting affective chatbot use may identify a particularly vulnerable population deserving much closer attention.
The broader trend is difficult to dismiss. Separate American research found nearly one in five adolescents and young adults had used AI chatbots for mental-health advice, while another youth study documented unsafe or troubling chatbot experiences. Together, the evidence suggests families and institutions cannot treat conversational AI as merely another educational technology.
A conservative response need not mean banning useful tools or handing government unlimited regulatory authority. It does mean recognizing that machines cannot assume the responsibilities of parents, counselors, teachers, clinicians, churches, families, and communities. AI can provide information, but emotional formation requires judgment, accountability, human relationships, and intervention when a child is genuinely in danger.
Sources
- https://jamanetwork.com/journals/jamapediatrics/article-abstract/2853453
- https://www.eurekalert.org/news-releases/1142125
- https://jamanetwork.com/journals/jamapediatrics/fullarticle/2849307
- https://onlinelibrary.wiley.com/doi/10.1002/jad.70164
- https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/