Perceived Impact of AI-Assisted Learning on Critical Thinking: Evidence From Chinese English Majors

This study explored how AI-assisted learning influences critical thinking skills (CTSs) among Chinese undergraduate English majors, focusing on analysis, evaluation, inference, and self-regulation from the Paul-Elder framework. Using a mixed-methods approach with 188 survey responses and 10 interviews, a tailored questionnaire measured students perceived CTSs changes. Findings show AI tools are widely used for writing, translation, and idea generation. While perceived analytical gains were modest (M=3.28), scores for evaluation, inference, and self-regulation hovered near neutral. Frequent AI use correlated negatively with self-regulation (r = -0.20, p = 0.007), suggesting risks of cognitive offloading. Interviews indicated that, despite recognizing AI’s limitations, students often relied uncritically on these tools under academic pressure and an efficiency-driven environment. The results highlight a core tension: as efficiency becomes central, independent thinking and metacognition risk being sidelined. Addressing how to shift AI from a cognitive shortcut to a true thinking aid emerges as a key educational challenge.