Cognitive Flexibility, Generative AI Adoption, and Problem-Solving Performance among University Students: The Mediating Role of Metacognitive Awareness

The rapid integration of generative artificial intelligence (GenAI) into higher education is reshaping how university students engage in learning, reasoning, and problem solving; however, the cognitive mechanisms underlying the effective use of these technologies remain insufficiently understood. This study examined the relationships among cognitive flexibility, GenAI adoption, metacognitive awareness, and problem-solving performance among university students, with particular emphasis on the mediating role of metacognitive awareness. A quantitative, cross-sectional research design was employed with a sample of 477 university students recruited from higher education institutions in Punjab and Sindh, Pakistan. Data were collected through an adopted and contextually adapted structured questionnaire administered online using Google Forms. The collected responses were systematically organized and analyzed using appropriate statistical procedures, including Pearson correlation analysis, multiple linear regression, one-way analysis of variance, and mediation analysis. The findings demonstrated significant positive relationships among cognitive flexibility, GenAI adoption, metacognitive awareness, and problem-solving performance. Cognitive flexibility, GenAI adoption, and metacognitive awareness also emerged as significant predictors of students’ problem-solving performance. Furthermore, students with higher levels of GenAI adoption demonstrated comparatively stronger problem-solving performance than those with lower levels of adoption. Mediation analysis further revealed that metacognitive awareness significantly mediated the relationships between cognitive flexibility and problem-solving performance and between GenAI adoption and problem-solving performance. These findings highlight the importance of students’ capacity to monitor, regulate, and evaluate their own cognitive processes when engaging with AI-generated information. The study contributes to the emerging literature on GenAI in higher education by demonstrating that the educational value of these technologies extends beyond technological adoption and is closely associated with students’ cognitive flexibility and metacognitive capabilities. The findings provide implications for educators, policymakers, and university administrators seeking to develop learner-centered and AI-integrated educational environments that strengthen independent thinking, reflective learning, adaptive cognition, and problem-solving capabilities while minimizing passive dependence on AI technologies.