The influence of college students' AI literacy on perceived usefulness: a moderated mediation model with AI anxiety and trust across disciplinary backgrounds.

BACKGROUND: To investigate how AI literacy influences perceived usefulness among university students from different disciplinary backgrounds, this study examines the roles of AI anxiety and trust. METHODS: Based on 563 valid samples collected from university students, a moderated mediation model was utilized to explore AI anxiety as a mediator and trust as a moderator across various disciplinary backgrounds. RESULTS: Results indicate that: (1) AI literacy significantly and positively predicted perceived usefulness, whereas AI anxiety negatively predicted perceived usefulness; (2) AI anxiety mediated the relationship between AI literacy and perceived usefulness; (3) Trust significantly moderated the relationship between AI literacy and AI anxiety, with high levels of trust buffering anxiety arising from low AI literacy; (4) Students from different disciplinary backgrounds exhibited differences in AI literacy, AI anxiety, trust, and perceived usefulness, yet trust's moderating effect remained consistent across disciplines. DISCUSSION: This study extends the Technology Acceptance Model (TAM) within an AI context, revealing the synergistic mechanisms of cognitive, affective, and socio-psychological factors in university students' AI adoption processes. The findings provide actionable guidance for AI education in higher education by emphasizing the need for discipline-specific curricula alongside targeted interventions that enhance trust and reduce AI anxiety, thereby improving students' perceived usefulness and engagement with AI technologies.