The differential role of AI-operated WCF in L2 students’ noticing of errors and its impact on writing scores

Abstract

Despite the plethora of studies on the role of noticing in second language learning, little is known about the role of AI-operated feedback in noticing errors and uptake during and after writing. To address this gap, this study primarily aimed to investigate the impact of feedback modes and timing on L2 students’ noticing of errors and writing scores. In pursuit of this objective, a quasi-experimental design was adopted in which 75 university undergraduate EFL participants were randomly distributed into four groups representing four feedback conditions. These included receiving feedback (a) from Grammarly while composing, (b) from E-rater after composing, (c) from the teacher after composing, and (d) a no feedback condition. Three argumentative prompts taken from IELTS task 2 were used as writing tasks and Noticing Sheets (NSs) were employed in recording and analyzing participants’ noticing. The findings generally demonstrate that participants exhibited increased noticing with teacher input compared to AI-operated feedback from Grammarly and E-rater, yet they still noticed errors when using the AI systems during and after writing. The results also show that noticing occurred in all feedback conditions, including the fourth one which involved no feedback from any source. This suggests that feedback is not an absolute prerequisite for noticing, but the source and timing of such corrective feedback can influence students’ noticing focus and intensity. The study provides further insights into the role of noticing in L2 learning and offers practical implications for teachers seeking to optimize conventional and automated feedback strategies for enhanced language learning.