VietTESOL International Convention 2026

This mixed-methods study explores the relationship between automated writing evaluation tools (AWE), such as ChatGPT, LexiCheck, and QuillBot, learner autonomy, and writing proficiency among 110 sophomore EFL students at a private university in Southern Vietnam. The participants were at an intermediate proficiency level and had primarily surface-level prior experience with AWE tools. The intervention lasted six weeks and was integrated into an IELTS writing course during February and March 2026. Both quantitative and qualitative data were collected, including pre- and post-intervention surveys examining students’ perceptions of AWE use, pre- and post-tests measuring students’ writing proficiency, and semi-structured interviews with selected students. Descriptive statistics of the survey responses, on writing process monitoring, goal setting, technology use, and proactive writing practice, revealed that learners generally rated their experiences positively (M ≈ 3.73–4.08), indicating that AWE significantly supported autonomous learning behaviors. Simultaneously, paired-samples t-test results for pre- and post-tests using IELTS writing test formats showed statistically significant improvements in all writing aspects (p < .001), including task achievement, coherence, vocabulary, and grammatical accuracy, with medium to large effect sizes (d ≈ 0.55–1.15), particularly in vocabulary and grammar. Furthermore, qualitative findings from interviews highlighted themes such as error detection, immediate feedback, improved autonomy, and skill enhancement. Additionally, students’ dependence on surface-level error correction and their level of active engagement were identified as key limitations. Overall, AWE improves writing ability and autonomy, and its efficacy is augmented by learners’ initiatives and approaches to using feedback.