Promoting Student Engagement in CSCL Through Scaffolds and Generative AI-Based Conversational Agent

Promoting student engagement in CSCL has been an important issue for educators. Previous studies have attempted to design retrieval-based conversational agents to assist students in completing collaborative tasks. However, few studies have explored how to integrate generative artificial intelligence with scaffolds to promote student engagement. Addressing the gap, this study developed a collaborative platform that integrates scaffolds with generative AI-based conversational agent. Additionally, a quasi-experiment involving 42 undergraduate students was conducted to examine the impact of this intervention on student engagement, including behavioral, cognitive, and emotional engagement. The results demonstrated that while the intervention did not significantly improve overall student engagement, it led to a notable increase in cognitive engagement. These findings suggest that generative AI-based agents, when combined with scaffolds, have the potential to target and enhance specific dimensions of engagement in CSCL environments.