Journal of affective disorders · 2026
Clinical and sociodemographic predictors of AI use for mental health among college students.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Liu Cindy H, Zhang Wenbo, Lou Felix, Zhao Chang, Chow Angela, Yip Tiffany
Аннотация
BACKGROUND: Generative AI tools are increasingly accessible to college students, yet little is known about who uses them for mental health support. This study examined predictors of AI use for mental health among college students at two U.S. institutions.
METHODS: Data were drawn from students (n = 896) who completed an AI module as part of the 2024-2025 Healthy Minds Study. The analytic sample comprised 675 students with complete data. Descriptive analyses compared three groups: never use AI, use AI but not for mental health, and use AI for mental health. Hierarchical logistic regression examined predictors of AI use for mental health using a binary outcome (AI use for mental health vs. no AI use for mental health), as the three-group structure could not accommodate general AI use as a predictor due to structural confounding. A supplementary multinomial logistic regression compared all three groups without general AI use.
RESULTS: Approximately 18% of students reported using AI for mental health. The never-use-AI group had higher proportions of non-binary/other gender and LGBQ+ students; the proportion of Asian students increased across groups in a stepwise pattern; and the AI-not-for-MH group showed better mental health profiles than both other groups. In regression models, frequent general AI use was the strongest predictor (OR = 11.42-12.87). 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 = 2.03-2.08). Lifetime therapy predicted AI use (OR = 2.21), but current therapy did not.
LIMITATIONS: Data were from only two institutions. The cross-sectional design precludes causal inference. Prevalence estimates are time-sensitive given rapid AI adoption.
CONCLUSIONS: Students with severe mental health symptoms and some marginalized groups are using unregulated AI tools at elevated rates. Findings underscore the need for research on AI safety for distressed individuals and policies accounting for heterogeneity in who uses these tools.
METHODS: Data were drawn from students (n = 896) who completed an AI module as part of the 2024-2025 Healthy Minds Study. The analytic sample comprised 675 students with complete data. Descriptive analyses compared three groups: never use AI, use AI but not for mental health, and use AI for mental health. Hierarchical logistic regression examined predictors of AI use for mental health using a binary outcome (AI use for mental health vs. no AI use for mental health), as the three-group structure could not accommodate general AI use as a predictor due to structural confounding. A supplementary multinomial logistic regression compared all three groups without general AI use.
RESULTS: Approximately 18% of students reported using AI for mental health. The never-use-AI group had higher proportions of non-binary/other gender and LGBQ+ students; the proportion of Asian students increased across groups in a stepwise pattern; and the AI-not-for-MH group showed better mental health profiles than both other groups. In regression models, frequent general AI use was the strongest predictor (OR = 11.42-12.87). 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 = 2.03-2.08). Lifetime therapy predicted AI use (OR = 2.21), but current therapy did not.
LIMITATIONS: Data were from only two institutions. The cross-sectional design precludes causal inference. Prevalence estimates are time-sensitive given rapid AI adoption.
CONCLUSIONS: Students with severe mental health symptoms and some marginalized groups are using unregulated AI tools at elevated rates. Findings underscore the need for research on AI safety for distressed individuals and policies accounting for heterogeneity in who uses these tools.