The Effects of AI Feedback Verification Training on Iraqi EFL Learners’ Writing Accuracy and Revision Quality

Abstract

This study investigated the impact of AI feedback verification training on Iraqi EFL learners’ writing accuracy and revision quality. Iraqi university EFL learners were assigned to an experimental group and a comparison group by means of an experimental design, quantitative pre-test-post-test. Both groups received the same kind of AI-generated feedback on their writing, but only the experimental group was explicitly trained to verify feedback by checking the intended meaning, consulting reliable language resources, and deciding whether to accept, modify, or reject AI suggestions. Linguistic accuracy was measured with standard indices of writing accuracy and revision quality was measured by examining the success of feedback-related revisions between the first and revised drafts. The results indicated that the AI feedback verification training significantly improved the learners’ writing accuracy compared to the traditional use of AI feedback without explicit verification training. The revision quality of the experimental group was also significantly higher. These results suggested that the effectiveness of AI feedback was not only about the availability of automated suggestions but also learners’ ability to interact with them critically and selectively. The study demonstrated the need for AI feedback verification, feedback literacy, and critical AI literacy in the EFL writing classroom, especially in contexts where generative AI is increasingly prevalent.