Who Benefits from AI Writing Tools? Rural Low-Income University Students’ Experiences of English Academic Writing and Educational Inequality

Abstract

This article examines how students from rural low-income backgrounds navigate support mediated by artificial intelligence (AI) in academic writing in English. Although generative AI is often framed as widening access to writing assistance, its equity implications remain insufficiently understood for students whose participation in academic English is shaped by rural schooling histories, economic constraints, and uneven digital resources. Drawing on a qualitative multiple-case study of 24 students and five teachers across two Chinese universities, the article analyses interviews, four-week writing journals, student writing artefacts, and institutional policy documents. The findings indicate that AI-mediated writing offered compensatory linguistic support: It reduced anxiety, improved surface-level accuracy, and helped students begin or revise writing tasks. These gains, however, were unevenly mediated by economic capital, cultural capital, and AI capital. The analysis also identifies a tension between more standardised academic English and students’ sense of academic voice. Integrating Bourdieu’s theory of capital with Freirean critical pedagogy, the article conceptualises AI-assisted writing as a contested educational space in which unequal advantage and more agentic forms of participation co-exist. More equitable provision requires transparent policy, explicit instruction in prompt design and output evaluation, and dialogic writing pedagogy that helps students retain ownership of their arguments and academic voice.