Reimagining Writing Feedback: The Effect of Copilot Artificial Intelligence (AI) on English Foreign Language (EFL) Students’ Written Performance

This study aimed to investigate the impact of using Copilot AI to provide corrective feedback on the written work of language learners and compare it with the corrective input from teachers. Using a quasi-experimental design, 70 Omani English language learners with an intermediate English proficiency level were assigned to a control group (which received general corrective feedback from the teacher) and an experimental group (which received detailed feedback from the Copilot AI and extra instructions form the teacher after Copilot AI), with 35 students in each group. Two comparison-contrast writing tasks were used as pretests and posttests and scored by two markers on Task Achievement, Organisation, Grammar and Lexicon. Although the initial findings revealed improvements in writing skills in both control and experimental groups, the detailed analysis of the posttest indicated that the experimental group performed significantly better than the control group in developing writing skills. In addition, the significant size effects of both tests revealed that both groups started with similar knowledge of writing skills in the pretest, but experimental group’s performance showed larger effect sizes which reveal higher performance in written tasks. The results of the study are insightful for teachers, learners and institutions.