Epistemic Agency and Academic Writing in the Age of genAI

Abstract

In higher education, academic writing has long been a central practice through which knowledge is developed, negotiated, and communicated. Epistemic agency, understood as responsibility for judgement and justification, has thereby functioned as a largely implicit foundation of writing, with authorship and text forming a stable unit from which epistemic responsibility could be inferred. The emergence of generative AI (genAI) disrupts this configuration. While genAI can participate in the formulation, structuring, and evaluation of knowledge claims, it does not bear responsibility for them. At the same time, it becomes internal to writing processes, making epistemic agency less visible and no longer reliably inferable from textual products alone. This article argues for a reorientation of academic writing pedagogy toward the explicit cultivation of epistemic agency under conditions of AI mediation. It conceptualises epistemic agency along four interrelated principles: decision-making, positioning, evaluation, and justification. From a media-pedagogical perspective, these are understood as mediated practices in which responsibility, critique, and authorship must be actively developed. The analysis thus situates academic writing within broader transformations of knowledge practices and points to fundamental challenges for learning and assessment in higher education in the age of generative AI.