AI-supported Collaborative Projects for Enhancing Problem-solving Skills in STEM Education

This study aimed to explore the role of Generative AI (GAI) in enhancing Collaborative Problem-Solving (CPS) capabilities within STEM education. To identify distinct types of learners based on CPS performance and to understand the interactions between learners and AI in the AI-supported projects, the research employed Epistemic Network Analysis (ENA) to analyze the cognitive networks of learners. Project-Based Learning (PBL) strategies was utilized to implement the AI-supported STEM collaborative projects and design teaching prompts procedurally in computer-supported collaborative learning environment (CSCL). As a result, the study identified three distinct clusters of learners with varying CPS performance. ENA results indicated varying interaction patterns and qualities among clusters, with implications for the effectiveness of AI as a collaborator in different stages of learning. In conclusion, the study concluded that GAI can significantly support CPS development in STEM projects, particularly when human collaboration is clear and directed. However, the effectiveness of AI is contingent upon the quality of directives provided by learners and the design of collaborative learning prompts. Future research should focus on in-depth study of prompt-design, and broadening the scope of factors considered in GAI-supported learning environments.