Beyond correct answers: epistemic agency and student learning in AI-assisted academic writing classrooms in Indonesia

Ridhatullah Assya’bani (Sekolah Tinggi Ilmu Al Quran (STIQ) Amuntai)

Indonesian Journal of Linguistics and Educational Studies
Vol. 1 No. 2 (2026): Generative AI, multilingual learning, and classroom divide in Indonesia

Background: The integration of generative artificial intelligence into higher education has transformed academic writing practices and raised important questions regarding students’ roles as active knowledge constructors rather than passive recipients of algorithmically generated information. Objective: This study examines how epistemic agency is represented and negotiated within AI-assisted academic writing environments in Indonesian higher education by focusing on the interrelationship between knowledge evaluation, human–AI collaboration, and AI literacy. Method: This study employed a qualitative corpus-based design using publicly accessible policy documents, institutional publications, academic articles, and educational media materials related to artificial intelligence, academic writing, and student learning published between 2023 and 2026. Results: Findings indicate that knowledge evaluation constitutes the most prominent dimension of epistemic engagement, suggesting that verification and source validation have become central learning practices in AI-mediated writing contexts. Findings further reveal that human–AI collaboration is predominantly framed through cognitive scaffolding functions, where artificial intelligence supports idea development, argument organization, and writing processes while students retain responsibility for final intellectual decisions. Implication: Findings also demonstrate that although AI literacy is increasingly associated with ethical awareness and critical evaluation, reflective engagement with the epistemological implications of AI remains comparatively underdeveloped across educational discourse. Novelty: The novelty of this study lies in its integration of epistemic agency, human–AI collaboration, and AI literacy within a single analytical framework, thereby shifting scholarly attention from questions of writing performance and technological adoption toward broader concerns regarding knowledge construction, intellectual responsibility, and learning in AI-assisted academic writing classrooms.