Investigating the mediating roles of learners motivation and expertise level in the relationship

Abstract

The rapid growth of generative AI tools like ChatGPT in universities has sparked ongoing discussions about how they affect students’ advanced thinking skills, such as problem-solving. Many people worry that these tools might reduce students’ ability to think critically, but there has been little clear evidence explaining how ChatGPT use actually influences learning. This study aims to fill that gap by testing a dual-mediation model based on an extended version of the Technology Acceptance Model (TAM). The model examines two key factors that may explain the link between ChatGPT use and problem-solving skills: learner motivation (an emotional pathway) and expertise level (a knowledge-based pathway). Using survey data from [N = 478] university students, the study analyzed results through Partial Least Squares Structural Equation Modeling (PLS-SEM). The findings strongly support the model. ChatGPT use was found to have a positive impact on problem-solving skills. Both motivation (β = 0.295, p < .001) and expertise (β = 0.153, p < .001) played significant roles in this process, with motivation showing a much stronger effect. Overall, the model explained 56.2% of the variation (R²) in problem-solving ability. In summary, this research helps to prove that AI tools like ChatGPT can improve learning. ChatGPT works both as a motivational tool that inspires students and as a cognitive aid that supports knowledge building. These insights can help educators design teaching strategies that use AI to improve, not replace, key learning skills. Research purpose: To examine how learners' motivation and expertise level mediate the relationship between ChatGPT usage and the development of problem-solving skills among university students. Research motivation: Despite considerable discourse on the impact of generative AI on student learning, there is a deficiency of research examining the specific mechanisms via which this influence manifests. This research is motivated by the imperative to transcend superficial direct impact evaluations and to empirically examine the underlying psychological and cognitive mechanisms at play. Research design, approach, and method: This study utilized a quantitative, cross-sectional survey design. A dual mediation model was suggested, based on the structure of the expanded Technology Acceptance Model (TAM). Data were gathered from a cohort of 478 university students and examined by Partial Least Squares Structural Equation Modeling (PLS-SEM). Main findings: The findings validate that the utilization of ChatGPT substantially enhances students' problem-solving abilities. This relationship is considerably and partially regulated by two distinct pathways: a psychological road through learners' motivation and a cognitive pathway through competence level. The model had considerable predictive efficacy, accounting for 56.2% of the variance in problemsolving abilities. Practical/managerial implications: The results give teachers and academic managers a more detailed picture of what they mean. ChatGPT is both a tool and a framework for learning. These insights can help people build educational programs that use AI tools to improve, not replace, the development of important problem-solving