Abstract
Artificial intelligence (AI) tools are rapidly reshaping how university students learn, yet the benefits of these tools may not be equally distributed. Grounded in Bourdieu's theory of capital and social reproduction and in digital capital theory (Ragnedda, 2018), this study examines whether digital capital shapes the perceived reproduction of educational inequality through the sequential mechanisms of AI literacy and AI usage for learning. A quantitative, cross-sectional survey was administered to 410 university students enrolled in undergraduate programmes in India. Exploratory Factor Analysis confirmed a clean four-factor structure (KMO = .886; Bartlett's χ²(190) = 3005.49, p < .001), and Confirmatory Factor Analysis in AMOS established satisfactory reliability and validity (Cronbach's α = .803–.846; composite reliability = .804–.846; discriminant validity confirmed via the Fornell–Larcker criterion and Heterotrait-Monotrait ratios, all below .60). Structural Equation Modelling revealed that Digital Capital significantly and positively predicted AI Literacy (β = .502, p < .001), which in turn predicted AI Usage for Learning (β = .581, p < .001), which significantly predicted Educational Inequality Perception (β = .529, p < .001). The direct path from Digital Capital to Educational Inequality Perception was non-significant (β = -.064, p = .292), while the standardized serial indirect effect through AI Literacy and AI Usage was significant (β = .154, 95% CI [.114, .206], p = .001), indicating full mediation. These findings suggest that unequal digital capital shapes perceptions of AI-driven educational inequality almost entirely through the literacy-usage pathway rather than through direct access alone, with implications for equitable AI integration in Indian higher education.