Developing EFL Learners’ Feedback Literacy through AI-Assisted Dialogic Interactions in English Academic Writing: A Mixed-Methods Study

Abstract

Feedback literacy—the understandings, capacities, and dispositions needed to make sense of feedback and use it While existing research has extensively examined AI-generated feedback in terms of writing product outcomes, the process of how learners develop feedback literacy through sustained interaction with AI remains insufficiently understood. Moreover, Carless and Boud’s (2018) framework was developed for human-to-human interaction, leaving open how its core dimensions translate into AI-mediated contexts and what AI-specific competencies might be required. To extend understanding in this area, we developed a GPT-4-based feedback bot, WriteWise, and engaged 45 Chinese EFL university students in a three-week mixed-methods experiment. We measured AI affordance for feedback literacy strategies by analyzing learner-bot conversations and evaluated effectiveness through pre-post-delayed feedback literacy tests, essay writing and revision tasks, and semi-structured interviews. Our findings suggest that AI afforded 14 sub-strategies across four main feedback literacy dimensions (Appreciating, Judging, Acting-on, and Generating feedback) plus one meta-level strategy. Appreciating and Acting-on strategies were most frequently used and significantly predicted feedback literacy knowledge and quality of English academic writing, though Appreciating strategies showed marginal significance (p < .10) for feedback literacy knowledge. However, Generating strategies (creating feedback for peers) were least utilized due to cognitive complexity and AI response limitations. We identified AI features influencing strategy frequency and effectiveness and offered implications for implementing AI-assisted feedback literacy instruction in English writing education.