Speaker
Description
As ChatGPT becomes increasingly common in EFL writing, understanding how students maintain active agency—rather than becoming passive editors of AI output—is critical. This study investigates how EFL students exercise agency when integrating ChatGPT into their writing processes. Drawing on learner agency theory and AI literacy frameworks, we examine students' decision-making patterns regarding AI-generated text, including how they plan, evaluate, revise, and reject suggestions. Data will be collected through AI interaction logs, student reflections, and semi-structured interviews within a class of 27 participants. The study compares writing patterns and agency levels between high-agency users and low-agency users. Findings will clarify mechanisms for fostering responsible AI integration in EFL instruction and assessment.
Keywords: AI agency; EFL writing; ChatGPT; learner autonomy