Interaction Patterns and Embodied Design Processes Using an AI Robotics Kit for K–12 Education

Abstract

This multidisciplinary study explores how children interact with a low-cost robotics educational tool designed to teach machine learning (ML) concepts through hands-on making. Artificial intelligence (AI) and ML are increasingly integrated into K-12 education, yet tangible approaches are not extensively studied. We conducted three researcher-led interventions with children aged 10, 11 and 13, each tasked to build an AI-enhanced robot using the robotics kit. Interventions were analyzed both qualitatively and with process mining methods to gain an in-depth view of the interactions. Findings indicate that designing and building were very central throughout the interventions, with the affordances guiding the design process forward. Children exhibited emerging data-driven thinking and ownership of their design process, while still requiring tailored scaffolding throughout the design process. We identify a general structure of the design process from all the interventions and emphasize the importance of building and designing activities. This study provides a new methodological lens to analyzing the interactions between children and embodied AI-educational tools while also contributing to the current body of knowledge on designing and effectively leveraging embodied learning tools.