When and How Students Engage with Augmented Generative AI During Mathematical Modeling Tasks

• In Event: Virtual Poster Session
  In Virtual iPresentation Only: SIG-Instructional Technology Virtual Poster Session

Abstract

This study investigates how students mobilize or discontinue their discursive practices with a retrieval-augmented generation AI model during community-based mathematical modeling tasks. The research team developed this model, named Gina, and adopted Cresswell’s mobility metaphors (routes, rhythms, and frictions) to analyze the dynamic nature of student–AI chatlogs. Findings show that AI prompts shaped students’ conversational pathways, supported routinized mathematical representations, and revealed moments of conversational breakdown. Using a socio-material lens, we highlight how student agency and cultural norms co-construct interactions with AI. This work expands current understandings of human–AI interaction by introducing a new conceptual framework and providing illustrative examples that inform both AI design and teaching practices.

Authors

• Eunhye Flavin, Georgia Institute of Technology

• Cristian Castaneda, Georgia State University