Effects of Different AI-driven Chatbot Feedback on Learning Outcomes and Brain Activity

Abstract Artificial intelligence (AI)-driven chatbots provide instant feedback to support learning. Yet, the impacts of different feedback on behavior and brain activation remain underexplored. We investigated how metacognitive, affective, and neutral feedback from an educational chatbot affected learning outcomes and brain activity through functional near-infrared spectroscopy technique. Students receiving metacognitive feedback showed higher transfer scores and metacognitive sensitivity, brain activation in the frontopolar area and middle temporal gyrus than other feedback types. Such activation correlated with metacognitive sensitivity. Students receiving affective feedback showed better retention scores than those receiving neutral feedback and higher activation in the supramarginal gyrus. Students receiving neutral feedback exhibited higher activation in the frontal eye fields and dorsolateral prefrontal cortex than other feedback types. The machine learning model outputs key brain regions that predict transfer scores. These findings underscore the potential of diverse feedback types in enhancing learning via human-chatbot interaction.