Skip to main content
Peer-reviewed Authoritative — Peer-reviewed venues, standards bodies, regulators, official government publications

Psychosis Risk and Generative Artificial Intelligence Use Frequency, Motivations, and Delusion-Like Experiences: Cross-Sectional Survey Study

Cross-sectional survey of 1,003 US young adults comparing GenAI users at elevated psychosis risk (Prodromal Questionnaire-Brief distress score of 20 or more) with low-risk users on frequency of use, motives, and chatbot interactions involving possible delusional content. Elevated-risk users were no more likely to have ever used GenAI but were more likely to use it intensively, to use it for social and emotional support, and to assign it human roles such as companion, therapist or romantic partner.

Read at jmir.org (opens the original source in a new tab) Archived copy(opens in a new tab)

Publisher

Journal of Medical Internet Research (JMIR Publications); University of North Carolina at Chapel Hill

Published

5 Mar 2026

Added

today

Key Findings

  • 267 of the GenAI users (28%) met the elevated-risk threshold and 685 (72%) did not
  • Elevated-risk users had higher odds of intensive use (OR 1.70 to 2.56 for several times a day, more than 30 minutes a day, or six or more conversations a day)
  • Elevated-risk users had higher odds of ascribing human-like roles to the chatbot: companion, friend, therapist or romantic partner (OR 1.76 to 3.08)
  • Delusion-related interactions were commonly endorsed in the at-risk group (item endorsement 13.3% to 30.7%)
  • Lifetime GenAI use did not differ between the risk groups

Methodology Notes

Online cross-sectional survey, n=1,003 US young adults. Psychosis risk measured with the PQ-B distress score (cut-off 20). Self-report and associational, so no causal direction can be drawn. JMIR landing page (jmir.org) returns a 202 shell. Verified via the Crossref record (created 2026-01-20; published online 2026-03-05; full abstract) and PubMed PMID 41785452 (abstract and UNC Chapel Hill affiliations).

Topics

Chatbot psychosisVulnerable usersDependency & parasocial attachmentHuman-AI relationshipsDigital mental health

Authors

Benjamin Buck, Anne Julia Maheux

Tags

psychosis-riskpq-byoung-adultssurvey

Cite This

APA

Benjamin Buck, Anne Julia Maheux. (2026). Psychosis Risk and Generative Artificial Intelligence Use Frequency, Motivations, and Delusion-Like Experiences: Cross-Sectional Survey Study. Journal of Medical Internet Research (JMIR Publications); University of North Carolina at Chapel Hill. https://www.jmir.org/2026/1/e85038