Tracing AI-Mediated Writing Tool Use Among Korean University Students: An Exploratory Screen-Recording Case Study
Received: Nov 14, 2025 ; Revised: Jan 31, 2026 ; Accepted: Mar 24, 2026
Published Online: Apr 30, 2026
ABSTRACT
This exploratory case study uses process-tracing to examine how four Korean EFL university students utilize online dictionaries, machine translation, and ChatGPT during English argumentative writing. Analyzing screen recordings across three drafts, the study identifies specific tool-use purposes, sequences, and frequencies. Results reveal that tool-use trajectories are independent of English writing proficiency, reflecting idiosyncratic strategic choices and personal styles. Despite its potential, ChatGPT use remained minimal and surface-level due to study constraints, exposing a gap between AI’s potential and actual application. As the instructional design prioritized restricting AI, participants’ usage patterns demonstrated a tendency toward safe and basic routines, hindering the development of strategic digital writing skills. The study advocates for guided support to foster student agency and reflective practice. Ultimately, this process-oriented analysis provides a framework for curriculum designers seeking to integrate generative AI in ways that are both academically responsible and cognitively expansive.





