A Study on the Factors Associated with College Students’ Intention to Use AI-Based Mental Health Services: The Roles of Trust, Empathy, and Privacy Concern

Artificial intelligence technology is gradually being integrated into mental health service contexts and is increasingly discussed as a potential supplementary resource to university mental health support systems. However, AI-based mental health services involve sensitive factors such as emotional expression, personal privacy, and relational interaction. Therefore, college students’ willingness to accept AI-based mental health services and the potential factors associated with such acceptance remain unclear. Based on an extended Technology Acceptance Model, this study introduces six factors—Accessibility (ACC), Privacy Concern (PC), Perceived Anthropomorphism (PA), Perceived Empathy (EMP), Trust (TRU), and Perceived Usefulness (PU)—to systematically examine their relationships with university students’ Behavioral Intention (BI) to use AI-based mental health services. A total of 375 volunteers participated in the current study and completed relevant questionnaires. Structural equation modeling was employed for data analysis. The results indicate that accessibility, perceived usefulness, and trust are all significantly and positively associated with behavioral intention. Perceived empathy is significantly and positively associated with trust, whereas privacy concern is significantly and negatively associated with both trust and behavioral intention. The analysis of indirect effects shows that there is a significant sequential indirect association between perceived anthropomorphism and behavioral intention through perceived empathy, trust, and perceived usefulness. In addition, a significant negative indirect association between privacy concern and behavioral intention through trust was observed. The findings indicate that, in the highly sensitive context of mental health services, college students’ intention to use AI-based mental health services is associated with a combination of accessibility, perceived usefulness, emotional interaction experience, privacy-related evaluations, and trust.