Generative AI and Educational Inequality

This article examines how generative artificial intelligence (AI) is accessed, used, and interpreted by repeat-year students in county-level high schools in China. Drawing on interviews, learning diaries, and AI interaction records from 24 students, the study explores how AI becomes educationally meaningful within a high-pressure and stratified learning environment. The article develops the Algorithmic Capital Reproduction Framework (ACRF), which combines Bourdieu’s theory of capital and social reproduction with a Freirean emphasis on critical engagement, critical consciousness, and the educational role of questioning unequal conditions. The findings show that while baseline access to AI tools was relatively widespread, meaningful and productive use remained uneven. Students differed in how they entered AI use, how they incorporated it into study practices, and how they evaluated and integrated AI-generated outputs. Interpretive competence emerged as a key mediator shaping whether AI supported reflective learning or shortcut-oriented task completion. Rather than reducing inequality, generative AI appears to reorganise it through socially differentiated patterns of access, engagement, and educational value.