<![CDATA[Generative artificial intelligence (AI) is transforming academic writing and challenging conventional understandings of authorship, linguistic competence, and academic voice. This study explores the linguistic characteristics of AI-generated and human-written academic texts and examines writers’ perceptions of authorship, linguistic voice, and AI-mediated writing. A qualitative case study involved 12 lecturers and postgraduate students in an Indonesian higher education context with experience writing academic texts in English and using generative AI. Data were collected through analyses of human-written and comparable AI-generated texts and semi-structured interviews. Thematic analysis using a corpus-informed qualitative approach focused on lexical, syntactic, discourse, metadiscourse, and authorial-voice patterns. Four themes emerged: negotiation of authorship, tension between linguistic smoothness and human voice, formulaic academic language, and AI’s shifting role from writing tool to writing partner. The findings suggest that linguistic sophistication does not necessarily indicate authorial authenticity. Instead, AI-mediated writing is better understood as a continuum of linguistic mediation rather than a binary distinction between human and machine production. The study connects corpus linguistics, metadiscourse, and authorial voice to conceptualize linguistic agency in AI-mediated academic discourse. It also highlights the need for AI literacy and writing pedagogy that prioritizes critical evaluation, rhetorical agency, and responsibility for academic language.]]>
Copyrights © 2026
