Artificial intelligence understanding to self-efficacy: The mediating roles of use behavior and skills

This study investigates how artificial intelligence understanding influences AI self-efficacy among university students within academic library contexts, with AI use behavior and AI skills examined as sequential mediators. Based on Bandura's social cognitive theory, self-efficacy is strengthened through conceptual understanding, which leads to purposeful AI use and subsequent skills acquisition. Using partial least squares structural equation modeling with data from 525 university students in Bangladesh, the results indicate that AI understanding significantly predicts AI use behavior (β = 0.777, p < .001), and AI use behavior significantly predicts AI skills (β = 0.573, p < .001). In turn, AI skills significantly predict AI self-efficacy (β = 0.468, p < .001), while AI understanding has a significant direct influence on AI self-efficacy (β = 0.268, p < .001). Mediation analysis further reveals a significant sequential indirect effect of AI understanding on AI self-efficacy through AI use behavior and AI skills (β = 0.440; 95% confidence interval = 0.338–0.542), indicating partial mediation.