From Exposure to Enactment: Distributed Accountability in AI Bias Analysis Tasks in English Writing for Academic Purposes

ABSTRACT The integration of generative artificial intelligence (AI) into academic writing has raised ethical questions about intellectual property, authorship, and knowledge representation. Although existing research has examined academic integrity and AI literacy, how learners negotiate responsibility when responding to biased or culturally inappropriate AI‐generated outputs remains underexplored. Grounded in critical applied linguistics and informed by relational accountability and distributed agency, this study examines how students enacted accountability in an AI bias analysis task within an undergraduate English academic writing course at a Chinese university. Drawing on group reports, AI interaction records, reflective writing, field notes, and interviews, thematic analysis identified three interrelated orientations through which students enacted accountability: epistemic accountability (evaluating knowledge claims), relational accountability (negotiating responsibility across human‐AI relations), and ecological accountability (questioning the broader linguistic, cultural, and epistemic consequences of AI‐generated knowledge). These orientations reveal accountability as distributed across epistemic judgments, social and technological relations, and broader conditions of knowledge representation. We conceptualize this configuration as distributed accountability: a situated process through which learners negotiate and enact answerability for AI‐mediated knowledge and its consequences. By demonstrating this expansion of answerability, the study connects the micro‐ethics of classroom practice with the macro‐ethics of linguistic and epistemic diversity and equity in AI‐mediated knowledge production. It offers implications for future research on academic writing ethics, critical AI literacy, and educational AI governance in applied linguistics.