AI-BASED PEER ASSESSMENT ON EFL STUDENTS’ WRITING

Abstract

Assessment is a fundamental component of education, serving not only to evaluate students’ learning achievement but also to provide meaningful feedback that fosters continuous improvement. Writing is often seen as one of the most difficult language skills in an EFL environment since it needs constant revision backed by excellent feedback. Although peer assessment has been widely implemented to promote collaborative learning and reflective practice, students often struggle to provide accurate, objective, and constructive feedback. The emergence of Artificial Intelligence (AI) offers new opportunities to enhance this process by assisting learners in generating more immediate and meaningful feedback during peer assessment. Therefore, this study was to ascertain how students' writing performance was affected by AI-Based Peer Assessment. 72 tenth-grade students were divided into an experimental group and a control group for a quantitative study using a quasi-experimental design. Pre-test and post-test writing assignments were used to assess students' writing proficiency. The data were analyzed using descriptive statistics, the Intraclass Correlation Coefficient (ICC) to establish inter-rater reliability, normality and homogeneity tests, paired-samples and independent-samples t-tests, and Cohen’s d effect size analysis. The findings revealed that students who participated in AI-Based Peer Assessment demonstrated significantly greater improvement than those receiving conventional peer assessment. Moreover, the large effect size indicated substantial practical benefits. These findings suggest that integrating AI into peer assessment is an effective strategy for enhancing students’ writing achievement through more accurate, timely, and constructive feedback