Arabic Academic Writing in the Age of GenAI: A Narrative Inquiry into Students Revision, Agency, and Dependency

his study examines how university students use generative artificial intelligence (GenAI) when revising Arabic academic texts, with particular attention to revision practices, learner agency, and dependency. Adopting a qualitative narrative inquiry design, the study draws on semi-structured interviews, revision notes, and AI-assisted revision records from eight undergraduate students in an Arabic Language and Literature program at an Indonesian Islamic university. The data were analyzed through thematic narrative analysis, moving from within-case reconstruction to cross-case interpretation. Four interrelated patterns emerged. First, students primarily used GenAI as a revision scaffold for grammar correction, lexical refinement, sentence reformulation, coherence, and polishing, rather than as a full-text generator. Second, their uptake of GenAI suggestions was negotiated through acceptance, modification, and rejection depending on meaning, tone, scope, and disciplinary fit. Third, agency remained visible through evaluative control, especially when students verified GenAI output against intended meaning, source fidelity, and authorial intention. Fourth, reliance on GenAI ranged from strategic support to emerging dependency, particularly when GenAI became a source of confidence, linguistic authority, and revision decision-making. Within this exploratory and context-bound study, GenAI-assisted revision is interpreted as a negotiated literacy practice rather than a simple case of technological assistance. By By foregrounding students’ narratives from one Indonesian Islamic university, the study offers a context-sensitive contribution to discussions of AI-assisted academic writing, academic literacy, and learner agency in AI-mediated environments. It further suggests that Arabic writing pedagogy should cultivate critical uptake, feedback literacy, and responsible GenAI-supported revision.