The Impact of ChatGPT-Generated Feedback on English Academic Writing Proficiency and Engagement: A Mixed-Methods Study

Generative artificial intelligence tools have emerged as practical solutions for providing automated corrective feedback that is immediate, scalable, and pedagogically aligned. While prior studies support the value of ChatGPT-generated feedback in improving surface-level English writing skills and fostering learner autonomy, few have examined its effects on broader academic writing dimensions and learner engagement through a mixed-methods lens. Addressing this gap, this study compared the impacts of ChatGPT-generated feedback and teacher feedback on the English academic writing proficiency and engagement of Chinese English-as-a-foreign-language learners enrolled in a foundation programme at a Malaysian university. Participants were assigned to an experimental group (EG), receiving iterative ChatGPT feedback, or a control group (CG), receiving rubric-based teacher feedback. Quantitative results showed significant improvement in both groups, with the CG outperforming the EG in grammatical and lexical complexity, while the EG demonstrated significantly greater gains in grammatical accuracy, coherence and cohesion, argument organisation, task achievement, and all dimensions of engagement (behavioural, cognitive, emotional). No significant difference was observed in lexical appropriacy. Qualitative findings of student journals and interviews from the EG revealed that ChatGPT feedback, perceived as prompt, non-judgmental, and rubric-driven, promoted autonomous revision and emotional safety. In contrast, teacher feedback was viewed by CG students as more linguistically helpful but emotionally pressuring or unclear. These findings suggest ChatGPT-generated feedback enhances rhetorical clarity and engagement, whereas teacher feedback supports deeper linguistic development. A blended feedback model that strategically combines both sources may best support academic writing instruction. Implications for pedagogy and future research directions are discussed.