AI-Assisted Feedback and EFL Writing Development: Evidence from Thai University Students in an English for Business Course

Abstract

This study examines the effects of ChatGPT-assisted feedback on the English writing development of Thai university students enrolled in an English for Business course. The research adopted a mixed-methods design involving 60 undergraduates divided into an AI-assisted experimental group (n = 30) and a teacher-led control group (n = 30). The participants were EFL students whose proficiency levels ranged from CEFR B1 to C1. Over an eight-week period, the experimental group used ChatGPT to support English writing and reflect on their own work. Pre- and post-test writing evaluations were analysed using t-tests, with significance set at p < 0.05. Participants’ reflections were analysed using thematic analysis. After triangulating the findings from the pre- and post-tests and reflections, the results showed that both the experimental and control groups significantly improved on their pre-test scores across all four writing criteria. The experimental group demonstrated substantially greater improvement than the control group across all four assessment criteria. More specifically, writing organisation showed the greatest improvement in the experimental group. The thematic analysis findings aligned with the quantitative data, indicating that the experimental participants became more aware of English language mechanics, organisation, and grammar. However, lexical development appeared to require more time for noticeable improvement. The study concludes that AI integration in ELT can serve as an effective tool for Thai learners in higher education. Further investigation into AI integration for EFL writing is recommended.