A community of inquiry perspective on human–AI co-facilitation within micro-blended learning

This study explores the distribution of teaching, social, and cognitive presences across human teachers and artificial intelligence (AI) systems in micro-blended learning environments for K–12 mathematics education, framed by the Community of Inquiry (CoI) framework. Utilizing a semester-long experimental design in high school Algebra classes, the instruction alternated between teacher-led in-person sessions and AI-facilitated online learning using MathPal, a conversational AI agent. The findings reveal a complementary relationship between teacher-led instruction and AI contributions, where teachers provided foundational conceptual scaffolds, affective support, and responsive guidance, while MathPal extended learning through step-by-step individualized scaffolding, adaptive feedback, and socio-emotional reinforcement. Sequential interaction analysis and hierarchical regression indicate that MathPal strengthened instructional continuity and engagement, while teacher-driven scaffolds remained the strongest predictor of mathematical reasoning outcomes. Based on these findings, this study introduces the Distributed Community of Inquiry framework, expanding CoI theory to address emerging dynamics in teacher-AI collaboration. Practical implications include optimizing modality alignment, embedding adaptive relational features into AI systems, and incorporating scalable learning designs that support equitable access to instructional support. This research contributes insights into the synergy between human and AI facilitators within micro-blended learning environments and highlights areas for future research, including broader validation across disciplines and AI platforms.