AUTHORIAL VOICE AND IDENTITY OF INDONESIAN EFL WRITERS IN AI-ASSISTED ACADEMIC TEXTS

Abstract

Authorial voice, the writer's recognizable presence in a text through stance, evaluation, and self-representation, is central to academic writing and to the identity of the academic writer. Generative artificial intelligence (AI) now participates in the production of academic texts, raising the concern that AI assistance may weaken, homogenize, or displace the voices of writers, particularly those writing in English as a foreign language (EFL). This systematic literature review (SLR), reported following PRISMA 2020, examined research on AI-assisted academic writing to determine what is known about the authorial voice and identity of EFL writers, with a focus on Indonesian writers. Drawing on a pool of 99 records and a supplementary set of 24 Indonesian records, 117 unique records were screened. No study met the core criterion of examining authorial voice or writer identity in AI-assisted academic texts, and none did so for Indonesian EFL writers. Forty-one studies were retained as adjacent evidence across four domains: authorship, authenticity, and agency (6 studies); the human–AI writing process (5); L2 and EFL writers and language (6); and the Indonesian context (24). The evidence is concentrated in the social dimension of voice (27 studies), followed by the authorial (15) and textual (9) dimensions. Synthesis suggests five themes: authorship is being redefined around accountability rather than production; human agency is treated as the guardian of voice; collaboration patterns determine how much of the writer remains in the text; L2 writers' critical AI literacy shapes whether they retain or surrender their voice; and Indonesian research frames AI use through ethics, policy, and practice rather than identity. The review proposes a layered model of authorial voice under AI and outlines a corpus-and-interview study design for Indonesian EFL writers.