The The Effect of Generative AI-Assisted Feedback on EFL Students’ Writing Performance

Abstract

The integration of generative Artificial Intelligence (AI) in education has introduced new possibilities for enhancing formative assessment practices, particularly in language learning contexts. In EFL classrooms, providing individualized and timely feedback on writing remains a persistent challenge due to teacher workload and limited instructional time. Generative AI systems offer potential support by producing immediate, structured feedback that may assist students during the revision process. This study investigates the effect of AI-assisted formative feedback on the writing performance of second-year baccalaureate students in an EFL context. Adopting a one-group pretest–posttest quasi-experimental design, 20 students completed an in-class writing task under controlled conditions. Their handwritten texts were evaluated using an analytic scoring rubric, then processed through a generative AI system to generate structured feedback focusing on grammar accuracy, vocabulary use, organization, and coherence. After reviewing the AI-assisted feedback, students revised their drafts, which were subsequently re-evaluated using the same rubric. Data analysis included descriptive statistics to examine mean score differences and a paired-samples t-test to determine whether observed improvements were statistically significant. The study provides empirical insight into the pedagogical value of AI-assisted formative feedback and contributes to ongoing discussions regarding the responsible integration of generative AI in EFL writing instruction.