Presentation AI-Assisted Writing Feedback in EFL: Tracking Student Performance and Reflections more

As automated writing assessment tools gain popularity in supporting independent learning, understanding their role in L2 writing development is essential. This study examines how Automated Essay Scoring (AES), specifically Write & Improve (W&I), supports students’ writing practice and influences their performance and perceptions in a Japanese university EFL course. Twenty-one students (B1 CEFR) completed nine writing tasks using W&I. A mixed-methods approach analyzed (1) relationships between task scores and final writing performance, (2) score differences between higher- and lower-level essay groups, and (3) patterns in students’ self-reflection comments. A Spearman’s rank-order correlation analysis showed no significant relationship between regular practice and final writing performance in either group. However, a Mann-Whitney U test revealed significant differences in narrative writing scores (p < .05), suggesting that essay genre had a more substantial impact on student performance than gradual writing development. Additionally, self-reflection data indicated that while many students found AI feedback helpful for identifying mechanical errors such as spelling, lower-level students grew frustrated when their scores did not improve despite repeated revisions. This presentation will discuss how AES tools can support L2 writing instruction, highlighting both their potential and the challenges of integrating more advanced AI feedback into writing practice.