Abstract
This mixed-methods study explores first-year EFL students’ perceptions and experiences of AI-supported learning in comparison with peer collaboration for critical reading. Specifically, it examines students’ perceived helpfulness of AI and peer support across critical reading processes and their preferences for these two learning modes, using a phenomenological stance to deeply explore these experiences. A total of 48 students participated in two learning sessions involving peer collaboration and AI-supported reading. Data were collected through a Likert-type questionnaire and semi-structured interviews. Wilcoxon signed-rank tests showed no statistically significant differences between AI support and peer collaboration in students’ perceived helpfulness across most critical reading processes, including identifying main ideas, analyzing arguments, detecting bias, and evaluating overall meaning. However, AI was perceived as significantly more helpful for understanding difficult vocabulary. Qualitative findings revealed complementary roles of the two learning modes: AI supported vocabulary processing, efficiency, and procedural assistance, whereas peer collaboration promoted diverse perspectives, shared meaning-making, and affective support. Students also expressed strong preference for combining AI and peer collaboration. Interpreted through the Technology Acceptance Model, the findings suggest that AI acceptance was shaped not only by perceived usefulness but also by trust in AI outputs and learners’ trust or confidence in their own critical judgment when interacting with AI for critical reading.