Why do students differ in AI competence? Examining the role of attitudes across academic pathways

Students’ attitudes toward Artificial Intelligence (AI) are increasingly recognized as critical drivers of engagement in AI-supported learning. However, the extent to which these attitudes translate into perceived AI competence remains insufficiently understood, particularly within differentiated curricular contexts. Drawing on an integrated Expectancy-Value Theory (EVT) and Opportunity-to-Learn (OTL) perspective, this study examines the relationship between students’ AI attitudes and perceived AI competence and investigates whether this relationship varies across academic tracks. Rather than conceptualizing academic tracks as opportunity-to-learn itself, academic tracks are treated as proxy indicators of differentiated curricular affordances, AI-related exposure, and learning opportunities. AI attitude is conceptualized as a multidimensional construct reflecting exploration, determination, and collaboration, while AI competence captures students’ perceived cognitive, ethical, and applied capabilities in engaging with AI systems. Survey data collected from school students across varied academic tracks were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM), complemented by Multi-Group Analysis (MGA), endogeneity assessment, and predictive evaluation using PLS-Predict. The findings show that students with more positive attitudes toward AI reported higher perceived AI competence, but this association varied across academic tracks. This suggests that attitudes alone do not translate into competence uniformly; rather, their association with perceived competence may vary according to the learning opportunities available within students’ curricular contexts. Robustness checks indicated no evidence of endogeneity bias and supported the model’s predictive relevance. In essence, the study contributes to AI education research by integrating EVT and OTL into a unified explanatory perspective and by positioning academic tracks as contextual proxies for opportunity-to-learn conditions that shape how students’ value-based engagement develops into AI competence.