The effect of artificial intelligence-based feedback on improving the grammatical accuracy of EFL students

Abstract

Background: Grammatical accuracy remains a major challenge in English as a Foreign Language (EFL) writing. Teacher feedback plays an essential role in writing instruction, yet it is often constrained by time, workload, and consistency. Artificial intelligence (AI)-based feedback tools, such as Grammarly and GPT-powered checkers, offer rapid, individualized, and iterative corrections, creating new opportunities in writing pedagogy.

Aim: This study aims to examine the effectiveness of AI-based feedback in improving the grammatical accuracy of EFL students’ writing. In addition, it compares the impact of AI feedback with teacher feedback on students’ accuracy and learning autonomy.

Method: A quasi-experimental design was employed with 60 EFL students divided into experimental and control groups. The experimental group received AI-based feedback, while the control group received teacher feedback. Data were collected through pre- and post-writing tests over six weeks. Statistical analysis was conducted using a paired-sample t-test with SPSS 26.

Results: The experimental group demonstrated a significant improvement in grammatical accuracy (M = 82.6, SD = 5.9) compared to the control group (M = 74.3, SD = 6.2), with p < 0.001. Findings also revealed that students who received AI feedback were more engaged in self-correction and attentive to errors.

Conclusion: AI-based feedback has been proven effective in enhancing grammatical accuracy in EFL writing while also fostering students’ learning autonomy. Although teacher feedback remains crucial for contextual and affective support, AI tools can serve as an effective complement to reduce workload and promote independent writing skills.