A cross-sectional study in JAMA Pediatrics found that seeking emotional support from generative AI was associated with psychological distress in young people even after accounting for loneliness and their sense of mattering.

Published: August 31, 2026, 10:20 p.m. PKT · Reporting cutoff: August 31, 2026, 9:45 p.m. PKT

What you need to know

  • The study examined affective use of generative AI—turning to it for emotional support rather than ordinary task assistance.
  • That use was associated with higher psychological distress among children and adolescents.
  • The association remained after researchers considered loneliness and “mattering,” or feeling valued by other people.
  • The findings suggest emotional AI use may help identify young people who need human support.
  • Because the study was cross-sectional, it cannot show whether AI use increased distress or distress prompted AI use.

A teenager asking a chatbot for homework help and a teenager asking it to relieve despair are both “using AI,” but they are not doing the same thing. New research argues that the purpose of the conversation matters.

The study identifies emotional-support use as a marker associated with psychological distress. It does not diagnose a child, measure the quality of a particular chatbot response or demonstrate that talking to AI harms mental health. Its value is narrower: clinicians, parents and educators should not treat all generative-AI use as a single behavior.

Functional use and emotional refuge are different signals

Generative AI can help summarize information, brainstorm, translate or explain schoolwork. Affective use occurs when a person turns to the system for reassurance, companionship, validation or help managing difficult feelings.

The researchers report that this affective pattern was independently associated with distress after considering loneliness and mattering. Mattering describes the sense that other people notice, value and depend on you. The result suggests that emotional AI use contains information not fully captured by those two measures.

That does not make the behavior proof of a mental-health condition. It makes it a possible prompt for a careful human conversation.

The direction of the relationship remains unknown

Cross-sectional research measures variables during the same general period. It can reveal that two patterns occur together, but it cannot establish which came first.

Several explanations remain possible. Distressed young people may seek an always-available, nonjudgmental outlet. AI interactions could sometimes reinforce withdrawal or dependence. A third factor—such as limited access to trusted adults or professional care—could influence both. The study cannot separate those pathways.

Readers should therefore reject the headline “AI causes distress in teens.” That conclusion was not tested.

How other headlines framed it

  • JAMA Pediatrics used the neutral research framing of affective generative-AI use and youth mental health.
  • The JAMA Network media summary described emotional-support use as a distinct marker of distress and urged clinical and digital-literacy frameworks to distinguish it from functional assistance.

What adults should—and should not—take from it

The evidence supports curiosity, not surveillance or punishment. A young person may choose a chatbot because it is immediate, private or easier than explaining a feeling face to face. Shaming that choice could close a route to real support.

At the same time, a conversational system is not a licensed clinician, cannot reliably assess every crisis and may generate incorrect or poorly calibrated advice. Emotional AI should not be presented as a substitute for qualified care or trusted human relationships.

Bottom line: seeking emotional support from AI may be a signal that a young person is struggling. The study does not establish that AI caused the struggle, and it does not justify diagnosing someone from their technology use.

Sources

  1. Vaillancourt et al., JAMA Pediatrics, August 31, 2026.
  2. JAMA Network study summary.
  3. Clinical context: adolescent health and generative AI risks and benefits.

Editorial disclosure: The lead image is an original concept illustration using a non-identifiable fictional subject; it is not a photograph of a study participant. This article provides research context, not medical advice. SciQuest received no payment for this coverage. To report a possible error, contact SciQuest.