The impact of generative AI on university students’ learning outcomes via Bloom’s taxonomy: a meta-analysis and pattern mining approach
Qi Xia, Peng Zhang, Wendan Huang, Thomas K. F. Chiu
Abstract
Higher education is facing a huge challenge with the development of Generative AI (GenAI). Many researchers have explored the effectiveness of GenAI in university students’ learning outcomes. However, despite its importance and the inconsistent results in existing research, a notable gap remains in systematically evaluating the impact of GenAI in higher education using a meta-analytic approach. To address this gap, this study investigates the impact of GenAI on university students’ learning outcomes through the lens of Bloom’s Taxonomy. The results demonstrate that GenAI is an effective tool for enhancing students’ lower-order, higher-order, and non-cognitive skills. Discipline is identified as a moderating factor in the application of GenAI to higher-order cognitive skills, and it is recommended to consider discipline characteristics and learning content when designing educational activities and GenAI applications. Additionally, the association rule mining analysis indicates that GenAI positively influences students’ lower-order cognitive skills when in-class interventions are conducted within a week. Meanwhile, small-group instruction in STEM disciplines is identified as a promising approach for cultivating students’ higher-order cognitive and non-cognitive skills. We also critically discuss positive benefits and exercise caution when interpreting the findings and implications. These findings help to shape policy development and instructional approaches in higher education.