Brains on Copy-Paste Mode? A Mixed-Methods Study on the Relationship between Generative Artificial Intelligence Use and Students’ Critical Thinking Skills

The rapid integration of generative artificial intelligence tools, such as ChatGPT, into higher education has raised important questions about their relationship with students' critical thinking (CT) skills. This study explored these associations among 398 undergraduate students using a mixed-methods design. Quantitatively, participants completed validated measures of critical thinking and a survey assessing frequency of AI use in academic contexts. Correlation analyses indicated a significant negative association between AI use and Analyzing/Organizing skills. Moderation analyses revealed that academic achievement moderated this relationship for Analyzing/Organizing and Evaluating, with higher achievers showing a stronger negative association. No significant effects emerged for gender and age on the remaining CT dimensions. Likewise, qualitative analyses of 200 open-ended responses identified three user profiles: (1) instrumental/uncritical, comprising younger low-achieving male students emphasizing AI speed and utility while reporting reduced verification effort; (2) stable/complementary, comprising older high-achieving students perceiving little change in their evaluative practices; and (3) ambivalent/skeptical-trusting, comprising women with medium achievement, divided between increased skepticism and greater reliance on AI outputs. Supplementary analyses linking Analyzing/Organizing and Evaluating scores to these profiles showed that lower development of both skills consistently aligned with the instrumental/uncritical profile. The interpretive consistency between both analytical strands suggests that the relationship between generative AI use and CT skills is nuanced and profile-dependent, limited in scope to specific dimensions and student groups rather than reflecting a broad negative impact, while highlighting the risks of uncritical reliance and the potential of AI as a complementary academic tool when used reflectively.