Cognitive Agility, Generative AI-Assisted Learning, and Problem-Solving Effectiveness among University Students: The Mediating Effect of Metacognitive Competence

Abstract

The rapid integration of Generative Artificial Intelligence (GenAI) into higher education is transforming how university students access information, engage with learning tasks, and approach complex academic problems, making it increasingly important to understand the cognitive mechanisms that support effective AI-assisted learning. The effects of cognitive agility, Generative AI-assisted learning, metacognitive competence, and problem solving with an emphasis on the mediating role of metacognitive competence on university students in Punjab and Sindh, Pakistan were analyzed. A quantitative, cross-sectional research design was utilized. 397 students were selected from universities using an adopted structured questionnaire filled in using the online Google Forms. Data was collected and analyzed in SPSS after being organized and screened in Microsoft Excel. Among the study variables, cognitive agility and Generative AI learning had a positive correlation which was statistically significant and explained metacognitive competence and problem solving. In terms of predictors, cognitive agility, Generative AI learning, and metacognitive competence had a positive correlation which was statistically significant and explained problem solving, with metacognitive competence having a more significant contribution. According to the mediation analysis, metacognitive competence was a significant mediator of the relation between cognitive agility and problem solving and between Generative AI learning and problem solving. This means that in terms of flexible thinking and learning enhanced by AI, planning, monitoring, evaluating and controlling one’s cognitive processes is a crucial factor in the effective resolution of problems. This study indicates that describing Generative AI as a teaching tool would not be enough. Teaching flexible thinking and planning, monitoring and controlling through metacognition is of equal importance. These findings can help higher education professionals, curriculum developers, and policy creators build responsible, reflective, and cognitively empowering AI-enabled learning environments in universities.