Speaker
Description
This study investigates how hybrid generative artificial intelligence (GenAI)–teacher feedback supports revision quality and feedback enactment in English for Academic Purposes (EAP) writing among B2-level undergraduate students. Although GenAI can provide immediate and detailed feedback, its suggestions may be generic, inaccurate, or adopted uncritically. In contrast, teacher feedback offers contextualized pedagogical guidance but is often limited by time and workload. To address these complementary strengths and limitations, the study examines an AI-first, teacher-follow-up feedback model in which students receive one round of fixed-prompt GenAI feedback before receiving teacher feedback on the same draft. Using an explanatory sequential mixed-methods design, approximately 50 B2-level undergraduate students will participate in a 12-week intervention involving six writing–feedback–revision cycles. Students will write academic argumentative essays, receive feedback from both sources, revise their drafts, and complete revision memos explaining their revision decisions. Quantitative data from Draft 1 and Draft 2 scores and a revision coding scheme will be used to examine changes in writing quality and the nature of students’ lower- and higher-level revisions. Qualitative data from AI interaction logs, feedback records, revision memos, and semi-structured interviews will explore how students interpret, accept, modify, combine, ignore, or reject feedback from GenAI and the teacher. The study is expected to clarify how students enact feedback during revision and how this process relates to revision quality. Findings may inform ethically grounded EAP feedback practices that combine technological efficiency with teacher oversight and promote students’ critical, independent revision skills.
