Measuring the Impact of ChatGPT Feedback on Saudi Undergraduate EFL Learners' Writing Skills

Abstract

As the conventional methods of instruction no longer seem to effectively conform to the taste of today's tech-savvy young generation, AI-powered tools are increasingly being incorporated in ESL/EFL classrooms for their efficiency, precise correction, and specific feedback. Despite multiple studies, the empowering role of ChatGPT in enhancing English writing skills, particularly in Saudi Arabia, remains largely unexplored. Therefore, this study employed a quasi-experimental design to compare and contrast the efficacy of ChatGPT with that of conventional feedback. While the students in the experimental group received feedback on their writing errors from ChatGPT for seven weeks, students in the control group received feedback from their teachers. To collect data, pre-and post-tests were carried out, which required respondents to draft tales based on pictures. It was revealed that the experimental group, which used an intervention, demonstrated fewer errors for all inflected morphemes than the control group, which used conventional instruction. These findings suggest that the use of AI tools like ChatGPT for language education courses serves as an effective supplement to traditional instructional methods, providing immediate feedback and personalized corrections that enhance instructional effectiveness.