Companionship and Control: Experiencing ChatGPT as Writer-Identity Work in L2 Academic Writing

Abstract

Abstract The rise of generative artificial intelligence (AI) is reshaping the landscape of academic writing. Existing research has largely emphasized functional aspects of tools such as ChatGPT (e.g., efficiency, accuracy, and plagiarism) while often overlooking how writers actually experience such tools. Little is known about the phenomenon of encountering ChatGPT as a second/foreign language (L2) academic writer seeking to establish an authorial voice and develop a professional identity. As such, this study explores in-depth how three Chinese international students enrolled in a UK MA TESOL (Pre-Service) program experienced ChatGPT as both advanced L2 writers and future teachers. Employing Interpretative Phenomenological Analysis (IPA), data were generated through semi-structured interviews about the participants’ prior experiences with ChatGPT, participant-produced recordings of a recent writing session, and stimulated recall sessions based on those recordings. Treating the data sources holistically for each participant and then comparing across participants disclosed three Group Experiential Themes: “Writing that Sounds Better but Means Less,” “Emotional Companionship and Psychological Safety,” and “Taming the Algorithm.” These findings suggest that ChatGPT can be experienced as a mediating presence in the cognitive, emotional, and identity work of multilingual writers. They also reveal a paradox: felt companionship that must be actively controlled. In this sense, the experience of ChatGPT can be a space where identity work occurs through the encounter (e.g., in resisting homogenization, asserting stance, and re-owning phrasing). For TESOL postgraduates, the tool then becomes intertwined with the development of both academic literacy and a professional, authorial voice.