AI-Generated Vs Teacher-Generated Emergent Feedback: Effects on EFL Learners’ Writing Feedback Literacy and Perceptions

This mixed-methods study explored the influence of AI- and teacher-generated emergent feedback on Iranian upper-intermediate EFL learners' writing feedback literacy and perceptions of both feedback modes. The study included ۱۱۸ upper-intermediate EFL learners in the last balanced analytic sample, with ۵۹ participants per feedback condition. The feedback literacy scale was held two times: before IELTS writing instruction and after ۱۰ sessions. Following an explanatory mixed-methods design, initially descriptive statistics, assumption checks, and mixed ANOVA were conducted.  Next, semi-structured interviews were administered to ۲۰ learners, and the results were analyzed using thematic analysis. The results of quantitative data analysis showed that writing feedback literacy enhanced during time. Moreover, when the scores over two measurement points were averaged, they were different across feedback conditions. However, the Time × Feedback Condition interaction was not statistically significant. Hence, no statistical evidence was given to support and indicate different enhancements across AI- and teacher-generated emergent feedback.  The qualitative results revealed different perceptions. Learners related AI feedback to receiving an instant response, to the same diagnosis more than once, and to monitoring independently. They associated teacher feedback with context-related explanations, conversation, and feeling supported. Since the study did not encompass a no-feedback comparison condition, the overall increase should be interpreted with caution and should not be ascribed to emergent feedback.

کلیدواژه ها:

Artificial intelligence ، EFL Learners ، Perception ، Sociocultural theory ، Emergent Feedback