From error correction to strategy support: AI-powered metacognitive scaffolding in assessing EFL writing performance and self-regulatory engagement

Abstract

The lack of individualized feedback may lead foreign language educators to think about giving feedback using a variety of artificial intelligence powered (AI-powered) tools, increasing access to automated assessment systems. Furthermore, metacognitive scaffolding guides students in planning, reflecting, monitoring, and assessing their language progress which can be adaptive through AI-driven tools like grammarly and write & improve that alter feedback. The purpose of this study is to investigate how AI-enhanced adaptive scaffolding delivered through automated feedback prompts and reflective questions can improve English as a foreign language (EFL) learners’ writing performance and self-regulatory engagement. Along with a self-report questionnaire, the study used a quantitative approach with experimental design. The study sample consisted of 120 EFL students. Participants were enrolled in a 6-week English writing course; they were randomly assigned to two groups: 60 students in the control group receiving conventional teacher feedback and 60 students in the experimental group receiving AI-assisted scaffolding via grammarly and write & improve. Data were collected through a self-report questionnaire on self-regulated learning strategies and writing tasks and analysed using ANCOVA to assess group differences. The findings show that students’ reflective practices, self-regulation, and personalized feedback were all strengthened by AI adaptive scaffolding. They suggest that AI-powered, adaptive, and metacognitively oriented feedback can extend the role of automated feedback from error correction to supporting learners’ strategic engagement with writing tasks. Implications are discussed for integrating AI-driven scaffolding into mainstream EFL pedagogy to enhance both assessment and self-regulated learning.