University Students' Writing Feedback Literacy in the AI Era: The Interplay of Generative AI Acceptance, Writing Anxiety and Writing Self‐Efficacy in Chinese EMI Educational Contexts

ABSTRACT As artificial intelligence (AI) tools reshape second language (L2) writing, how technological and psychological factors jointly shape English‐medium instruction (EMI) students' writing feedback literacy remains underexplored. Using an online questionnaire with 500 Chinese EMI university students, this study employed structural equation modelling (SEM) to examine how generative AI acceptance, writing anxiety and writing self‐efficacy shape students' writing feedback literacy. The results showed that (1) generative AI acceptance positively and significantly predicts students' writing feedback literacy and writing self‐efficacy, and negatively predicts writing anxiety; (2) writing self‐efficacy positively mediated the relationship between generative AI acceptance and writing feedback literacy, while writing anxiety mediated the relationship by reducing negative emotions; and (3) generative AI acceptance improves writing feedback literacy by raising students' self‐efficacy and lowering their writing anxiety. This study broadens understanding of writing feedback literacy in the AI era by clarifying the psychological mechanism linking generative AI acceptance to writing feedback literacy in EMI writing, and offers pedagogical implications for integrating AI‐based support to improve EMI students' emotions and feedback engagement.