How does artificial intelligence literacy affect university students’ innovative behavior? A serial mediation analysis based on latent profiles

Purpose: Against the backdrop of the rapid proliferation of generative artificial intelligence (AI), how university students can effectively leverage AI technologies to foster their innovative behavior has emerged as a critical issue in higher education. This study aims to explore the relationship and underlying mechanisms between AI literacy and university students' innovative behavior. Methods: A cross-sectional survey was conducted among 1,040 university students in China. Data were collected using the AI Literacy (AIL) Scale, AI Anxiety (AIA) Scale, Cognitive Flexibility (CF) Scale, and Innovative Behavior (IB) Scale. Results: < 0.001). Mediation analysis revealed three significant indirect pathways: the independent mediating effect of AIA accounted for 13.44% of the total effect, the independent mediating effect of CF accounted for 14.15%, and their serial mediating effect accounted for 3.30%. Additionally, latent profile analysis identified three latent profiles of AI literacy: low AI literacy (15.29%), moderate AI literacy (37.02%), and high AI literacy (47.69%). Significant differences were observed in the serial mediation pathways across these distinct profiles. Conclusion: AIL was directly associated with university students' IB and showed indirect associations through the independent and serial mediation of AIA and CF. These findings extend the application of affordance theory in human-AI collaboration contexts and provide empirical evidence for higher education institutions to implement differentiated AI education and innovation cultivation strategies.