Integrating ChatGPT into 5E inquiry-based learning to scaffold EFL academic research writing

Abstract

Large language models such as ChatGPT are increasingly used in education, yet evidence on how to integrate them systematically into EFL academic writing pedagogy remains limited. This study examined a 16-week undergraduate EFL composition course at a national university in Taiwan that embedded ChatGPT within the 5E inquiry-based learning (IBL) cycle (Engage, Explore, Explain, Elaborate, Evaluate). Students completed sequenced research-writing tasks (outline, questionnaire design, initial findings, abstract/introduction, and a full research report) with ChatGPT used as a dialogic scaffold. Using an exploratory classroom-based mixed-methods design, the study collected pretest and posttest writing scores, peer-discussion transcripts, writing artifacts, ChatGPT interaction logs, and reflective journals. Descriptive statistics and paired-samples t-tests examined score change, while qualitative content and thematic analyses triangulated engagement patterns across phases. Results showed a modest, non-significant increase in timed argumentative writing performance, while qualitative findings suggested process-oriented development in inquiry-based academic research writing practices. Students used ChatGPT throughout the 5E cycle for higher-order inquiry work, including topic framing, instrument refinement, and data interpretation as well as for drafting and language polishing. Across the semester, they developed phase-sensitive prompt literacy and critical filtering by selectively adopting, revising, or rejecting AI suggestions to maintain authorial control. An illustrative case further demonstrated how this phase-sensitive ChatGPT engagement unfolded across the five inquiry phases. Overall, the 5E IBL structure appeared to channel AI use toward purposeful inquiry and responsible academic writing development.