Abstract
OBJECTIVE: Chatbots' rapid advancements raise the possibility that they can be used to deliver mental health support. However, public utilization of and opinions toward chatbots for mental health support are poorly understood. METHODS: Survey study of 428 U.S. university students who participated in early 2024, just over one year after the release of ChatGPT. Descriptive analyses examined utilization of and attitudes toward both traditional mental health services (i.e. psychotherapy, counseling, or medication) and chatbot-delivered mental health support. RESULTS: Nearly half (49%) of participants reported having used a chatbot for any purpose, yet only 5% reported seeking mental health support from a chatbot (8% when only considering participants with probable depression or generalized anxiety disorder). Attitudes toward traditional mental health services were broadly positive, and attitudes toward chatbot-delivered support were neutral and significantly less positive (d = 1.18, p < .001). Participants reported lack of need and doubts about helpfulness as barriers to using chatbot-delivered support more frequently than they reported them as barriers to traditional services. Cost, time, and stigma barriers were less frequently reported for chatbot-delivered support than for traditional services. Attitudes were generally consistent as a function of mental health status. CONCLUSION: Among U.S. students, utilization of chatbots for mental health support is uncommon. Chatbots are perceived as less likely to be beneficial, yet also less affected by cost, time, and stigma barriers than traditional services. Rigorous outcome research may increase public trust in and utilization of chatbots for mental health support.