From Passive Answers to Active Inquiry: How AI Supports Critical Reading in EFL Classrooms

This study investigates the integration of ChatGPT, a generative AI tool, into critical reading instruction for university-level EFL learners. Recognizing the importance of higher-order reading skills such as evaluating arguments, recognizing bias, synthesizing information, and generating counterarguments, the research explores how AI-supported tasks influence both skill development and learner perceptions. Using a mixed-methods sequential explanatory design, the study engaged 35 second-year English majors in an eight-week intervention. Data collection included pre- and post-tests, five AI-integrated reading tasks, a structured questionnaire, and semi-structured interviews. Quantitative results revealed significant improvements in students’ ability to recognize bias, generate counterarguments, and identify main ideas. Qualitative analysis of ChatGPT interaction screenshots and student reflections demonstrated behavioral progression from surface-level clarification-seeking to more critical inquiry and evaluative questioning. Students reported high motivation and appreciation for AI's role in supporting comprehension and analysis. However, they also expressed skepticism, frequently verifying AI outputs and reflecting on potential bias. These behaviors contrast with earlier studies that emphasized uncritical reliance on AI tools, highlighting the emergence of critical digital literacy in this context. The findings suggest that AI can serve not only as a comprehension aid but also as a scaffold for cognitive and ethical engagement. When paired with explicit instruction, AI tools hold promises for fostering critical literacy and responsible use of technology in EFL education.