The Impact of AI Tools on IELTS Writing Preparation: Feedback, Autonomy, and Learner Perceptions

This mixed-methods study investigates the integration of Artificial Intelligence (AI) tools in IELTS Writing Task 2 preparation among EFL learners (n = 18). Employing a descriptive cross-sectional survey design, the research combines quantitative Likert-scale data with qualitative thematic analysis to examine learner perceptions of AI-mediated feedback, idea generation, and the associated risks of cognitive dependency. The findings reveal a dualistic relationship with technology: while students demonstrate strong reliance on AI for surface-level linguistic accuracy (grammar correction: M = 3.56; vocabulary enhancement: 77.8% usage), they remain skeptical about AI's capacity to replace human pedagogical guidance (M = 2.94 for AI as a teacher substitute). Qualitative data further exposes critical tensions, including inconsistent scoring reliability, linguistic over-complexity, and the phenomenon of "cognitive inertia"—a psychological dependency that undermines learner autonomy. Despite these challenges, AI emerges as a valuable scaffold for autonomous practice, eliminating feedback waiting periods and providing immediate linguistic support. The study concludes that AI functions optimally as a complementary "micro-level" editor rather than a substitute for human instruction, proposing a hybrid pedagogical model that integrates technological efficiency with critical human mentorship. Recommendations for students, educators, and institutions emphasize the development of AI literacy and structured reflection to preserve authentic intellectual development in high-stakes writing contexts.