Prevalence and correlates of GenerativeAI chatbot use for mental health concerns in higher education students

Background Large language model chatbots are increasingly used for mental health help-seeking, with general-purpose systems like ChatGPT becoming popular among young adults despite concerns about accuracy and misinformation. High rates of mental distress in higher education attending emerging adults make them an important focal population to characterize and contextual the use of generativeAI models for mental health in. Methods Data were collected from the 2024-2025 Healthy Minds Study at a large, public 4-year university. Of 12,000 students invited, 2071 participated; 896 bachelor-level students aged 18–24 years were randomly selected to complete a module with generativeAI questions, and 809 (90.2%) had complete data. Measures included chatbot use for mental health support, perceived AI trustworthiness, AI knowledge, mental health diagnosis, perceived need for help, past-year mental health treatment, barriers to care, other forms of informal help-seeking, and demographics. Weighted logistic regression models examined associations between these variables and chatbot use for mental health support. Results 15.4% of students reported using generalized chatbots for mental health help-seeking. Significant correlates included greater perceived need for help, experiencing more barriers to clinical care use, higher perceived trust in AI for mental health information, lower AI knowledge, and identifying as Asian. Specific barriers of financial constraints, limited time, fear of clinician mistreatment due to identity and privacy concerns were linked to chatbot usage. Conclusions GenerativeAI chatbots are a notable source of mental health support for college students, especially among those facing access barriers and who perceive AI as trustworthy. Future initiatives should educate young adults about chatbot capabilities and limitations and promote equitable mental health support.