An AI-assisted critical thinking intervention to enhance undergraduate EFL learners’ writing proficiency

Abstract

This study investigates the effects of an AI-assisted critical thinking (CT)-oriented writing intervention supported with ChatGPT involving 250 undergraduate EFL learners from three public universities. The research measured changes in students’ writing proficiency and their acceptance of ChatGPT in CT-oriented writing practices. A pretest-posttest design was adopted, with data collected through pre-post-writing tests (n = 250), an acceptance questionnaire (n = 250), and further semi-structured interviews (n = 25). The writing tests were evaluated across nine dimensions: clarity, accuracy, precision, relevance, depth, breadth, logic, significance, and fairness. The acceptance questionnaire, grounded in the Extended Unified Theory of Acceptance and Use of Technology (UTAUT2) model, assessed six constructs, i.e., performance expectancy, effort expectancy, hedonic motivation, habit, trust, and behavioral intention. The results showed that 1) The intervention significantly enhanced students’ CT reflected in writing; 2) Positive shifts were observed in four scales of ChatGPT acceptance in a CT-oriented writing context, i.e., performance expectancy, effort expectancy, trust, and behavior intention, though only marginal increase in hedonic motivation and habit. Meanwhile, the study identified the primary aspects of ChatGPT influencing EFL learners’ CT-oriented writing. These results demonstrate the potential of ChatGPT as a scaffolding tool for CT-oriented writing training in EFL contexts. Despite limitations such as the single-group experimental design and 8-week experimental period, the findings offer valuable insights for further research and development in AI-driven tools in EFL writing pedagogy. Future research could adopt a multi-group design and longitudinal assessments to further validate the intervention’s sustained effects.