Is Gen AI an editor or a revision tool? A case study of EFL students in AI-assisted revision

As generative artificial intelligence (Gen AI) tools become increasingly integrated into language education, understanding how learners respond to AI-generated feedback during writing revision has become an important research concern. Existing studies have primarily examined the effects of AI on writing performance, with less attention to learners' reported decision-making and reflection during revision. This qualitative study investigated learner-AI interaction, reported revision practices and learner perceptions among 18 first-year non-English-major students enrolled in a university EFL writing course. Data were collected through semi-structured interviews following a Gen AI-assisted revision task and analysed using thematic analysis. The findings identified three interaction patterns with Gen AI: passive AI reliance, adaptive AI engagement, metacognitive AI engagement. Participants differed in how they evaluated and used AI-generated feedback, including passive acceptance, selective uptake, and critical evaluation. Some reported increased confidence, motivation, and awareness of linguistic form, whereas others expressed concerns about overreliance, authorship, and loss of personal voice. These self-reported experiences provide insight into how EFL learners perceive and manage GenAI-assisted revision, highlighting the need for pedagogical approaches that foster critical AI literacy and reflective learner autonomy.