Smarter AI, Healthier Students? How Perceived AI Assistant Intelligence Shapes Medical Students’ Mental Health Through Learning Goal Progress and Academic Anxiety

As artificial intelligence (AI) adoption grows in medical education, AI assistants have become important learning support tools for medical students. However, existing research has primarily focused on learning performance and technology adoption, while little is known about how perceived AI assistant intelligence influences medical students’ mental health. Drawing on Self-Determination Theory (SDT), this study develops a model linking perceived AI assistant intelligence to mental health through the serial mediating roles of learning goal progress and academic anxiety, while examining the moderating role of AI literacy. A two-study design was employed. Study 1 surveyed 721 Chinese medical students, and Study 2 conducted a real human–AI interaction experiment with 398 medical students. Results showed that perceived AI assistant intelligence positively predicts learning goal progress. Learning goal progress and academic anxiety jointly mediate the relationship between perceived AI assistant intelligence and mental health. In addition, AI literacy strengthens the positive effect of perceived AI assistant intelligence on learning goal progress and enhances its indirect effect on mental health through the serial mediation pathway. The experimental findings further support the robustness of these relationships. This study extends research on AI in education by uncovering the psychological mechanisms through which perceived AI assistant intelligence affects medical students’ mental health. It also enriches the application of SDT in human–AI collaborative learning contexts and provides practical implications for optimizing AI-assisted learning environments, improving AI literacy, and promoting student well-being.