AI and the Developing Child: Reimagining Social-Emotional Learning Through Interactive Media Characters

• In Event: Teaching and Learning with AI: Building Literacy, Agency, and Design.

Abstract

This study explores how varying levels of interactivity with an AI media character affect young children’s social-emotional engagement (SEE) and learning outcomes. A quasi-experimental design assigned 240 children aged 3–6 to one of three conditions: Conversational AI, Pseudo-Interaction, or No Interaction. Children watched a PBS Kids episode featuring Elinor, a interactive AI chatbot character. Then SEE was measured through trust, enjoyment, and relational closeness, while learning outcome was evaluated using a structured comprehension post-test. Results showed children in the Conversational AI condition demonstrated modest but consistent gains in both SEE and learning outcomes. Findings reimagine educational futures where emotionally responsive AI may foster relational and developmentally age appropriate learning experiences for young children and inform future classroom instructions and curriculums.

Authors

• Peixin Zhou, Harvard University

• Ying Xu, Harvard University