Writing with the Machine: Negotiating Authorial Identity in AI-Assisted Academic Writing Practices among Post-Graduate Students

Abstract

Generative AI (GenAI) has revolutionized academic writing, shifting the nature of human-AI interactions in text production. While previous studies focus on academic honesty and general writing outcomes, little research addresses how multilingual university students negotiate authorial identities during AI integration. This qualitative study explores the negotiation of authorial identity among 15 postgraduate students at Mahendra Multiple Campus, Tribhuvan University, Nepal, using semi-structured interviews and reflexive thematic analysis. Guided by Ivanič’s Writer Identity Theory, Vygotsky’s Sociocultural Theory, Norton’s Identity and Investment Theory, and Bourdieu’s Linguistic Capital Theory, the study investigates students' perceptions of authorship, creativity, and linguistic voice. Empirically, the findings reveal five key outcomes: (1) students utilize GenAI primarily as a linguistic scaffolding tool to bridge conceptual ideas and formal academic English; (2) authorship is experienced as a negotiated sociotechnical space where intellectual ownership of ideas remains human while linguistic expression is co-constructed; (3) GenAI acts as a cognitive and creative partner that stimulates alternative argumentation rather than replacing human ingenuity; (4) participants experience an internal tension between achieving polished Western-centric academic English and preserving their authentic multilingual voices; and (5) sustainable writing relies on robust human agency, critical evaluation, and editorial oversight. These findings demonstrate that AI is not a mere substitution tool but an intervening mediator that reshapes scholarly writing. The study offers critical insights from a South Asian multilingual higher education context, advocating for AI-literate classrooms that foster intellectual ownership while enabling responsible technological utilization.