Students’ Experiences of Generative AI in Higher Education

Understanding Digital Inequalities Through a Sociomaterial Lens

Abstract

The integration of artificial intelligence (AI) in education has been widely discussed in recent years, with many studies emphasizing its potential to enhance learning efficiency and transform traditional educational practices. The emergence of generative AI (GenAI) systems such as ChatGPT, Claude, and Gemini has further accelerated these conversations, highlighting both the promise and the disruption such technologies bring to higher education. However, much of this discussion remains technology-driven, focusing primarily on what AI can do to improve learning rather than how students actually experience and appropriate these tools in their daily study practices. This short paper addresses this gap by examining how GenAI engagement is shaped by uneven conditions of participation across institutional contexts. Drawing on sociomaterial perspectives and postdigital critiques of “digital divide” thinking, the study frames GenAI-related inequality not as a simple access gap, but as emerging through entanglements of tools, infrastructures, institutional norms, and social support. Little empirical work has examined how differences in tool availability, guidance, and institutional arrangements shape students’ ability to engage meaningfully with GenAI, beyond individual “literacy” or willingness to use it. By comparing students’ experiences across university types, this paper shows how GenAI-related inequalities are enacted through everyday practices, including access pathways, support networks, and local norms around legitimate use.

This short paper reports findings from an ongoing study of students’ day-to-day digital practices across two different types of universities in China. Drawing on sociomaterial perspectives, it treats students’ engagement with GenAI not as an isolated or purely cognitive activity, but as shaped through dynamic interactions among both human and non-human actors. Using qualitative interview data from students at a highly ranked university and a less prestigious university in East China, the paper examines how access conditions, forms of guidance, and institutional cultures shape engagement with AI tools. The findings show that students’ engagement with GenAI is assembled through different networks of resources and constraints. In the elite university, wider tool availability, stronger infrastructural support, and more established peer and institutional guidance enable more sustained and strategic use of GenAI for academic work. In the less prestigious university, constrained access pathways, limited guidance, and greater uncertainty about legitimate use restrict how confidently and effectively students can incorporate GenAI into study routines. These contrasts demonstrate how GenAI-related inequalities are produced through everyday participation conditions rather than access alone.