Presentation
Exploring Artificial Intelligence Tutor Adaptability to Harness Discovery Curiosity and Promote Learning in Applied Mathematics and Life Sciences
SessionPoster Session 2
DescriptionAs Artificial Intelligence (AI) increasingly enters educational domains, its role as a dynamic teammate rather than a static tool demands closer examination. This study explores how an AI teammate powered by a Large Language Model can foster curiosity-driven engagement and learning through real-time adaptive guidance during Interactive Molecular Dynamics (IMD) tasks on the Visual Molecular Dynamics platform. Over four structured 15-minute sessions, high school students completed IMD tasks of complex molecular visualization and calculations. The AI employed curiosity-triggering behaviors (posing questions) and curiosity-response behaviors (providing detailed explanations), forming a feedback loop to sustain curiosity and deepen understanding. Performance was scored across three dimensions—task completion, conceptual understanding, and engagement—and teams were grouped by performance level based on a composite score. Communication patterns were analyzed using Cross Recurrence Quantification Analysis, which captured coordination metrics, like recurrence rate, to assess coordination dynamics. Student questions were coded by complexity (basic, intermediate, advanced), while AI messages were coded as either triggers or responses. High-performing teams showed stronger synchronization, asked more advanced questions, and received more curiosity-triggering interventions. These findings underscore the importance of adaptive, structured engagement in human-AI teaming and provide design insights for intelligent educational systems that foster curiosity and enhance cognitive performance.
Contributors
Event Type
Poster
TimeWednesday, October 15th5:30pm - 6:30pm CDT
LocationRiverside East