Exploring the Influence of AI-Assisted Feedback on EFL Students’ Writing Performance and Self-Editing Skills

Abstract

This study explores the influence of AI-assisted feedback tools, specifically Grammarly, on EFL students’ writing performance and self-editing skills in a university context. Grounded in Cognitive Process Theory of Writing (CPTW), this research adopts a mixed-method design. It combines both a quantitative analysis of pre- and post-tests and qualitative data from surveys and interviews. Sixty contributors were split into two cohorts: a traditional one receiving teacher feedback, and an experimental one using Grammarly during drafting and revision. Results indicate that the AI feedback students substantially outperformed their counterparts in the control group in writing quality, particularly in grammar, lexical range, and content organization. Additionally, students using AI-assisted tools demonstrated more frequent and strategic self-editing behaviors and reported increased autonomy and confidence. This study empirically bridges AI feedback tools with CPTW, revealing how they foster iterative self-editing and autonomy, a gap in EFL writing research. While teacher feedback remains essential for higher-order concerns, the integration of AI tools supports iterative revision and independent learning. This investigation emphasizes the educational value of mixing AI with conventional instruction and suggests implications for designing curriculum and developing instructional tools.