Clinical and sociodemographic predictors of AI use for mental health among college students

Key Findings

• — Approximately 18% of students reported using AI for mental health
• — Frequent general AI use was the strongest predictor (OR 11.42 to 12.87)
• — Moderate depression (OR 2.06), severe depression (OR 2.49), severe anxiety (OR 2.04) and suicidality (OR 1.97) each predicted AI use for mental health
• — Asian students showed elevated odds (OR 2.03 to 2.08); lifetime therapy predicted use (OR 2.21) but current therapy did not
• — The never-use-AI group had higher proportions of non-binary or other-gender and LGBQ+ students, and the group using AI but not for mental health had better mental-health profiles than either other group
• — The authors conclude that students with severe symptoms and some marginalised groups are using unregulated AI tools at elevated rates

Methodology Notes

Journal of Affective Disorders, DOI 10.1016/j.jad.2026.122058, online 2026-05-29 (PubMed ArticleDate; PMID 42217639), November 2026 issue. Affiliations from PubMed: Brigham and Women's Hospital Pediatrics and Psychiatry, Harvard Medical School, Vanderbilt University, Indiana University School of Public Health, Fordham University Psychology. Cross-sectional data from two institutions in the Healthy Minds Study; the three-group structure could not accommodate general AI use as a predictor, so the main model is binary with a supplementary multinomial model. ScienceDirect blocks this box; the structured abstract was read via PubMed, full text not read.

Sources