Learning via AI Dolls: Creating Self-Active Learning for Children

Abstract

This research aimed to propose an artificial intelligence (AI) chatbot mobile application for pre-school children engaging in active learning processes. The research tools Microsoft Bot Framework and Azure Bot Service were used to create an AI chatbot doll (AIBD) prototype. With this AIBD, the players simply dragged and dropped items in an intelligent bot builder that could create characters with different identities using a set of customized doll items, for example, genders, dresses, shoes, or flowers. During play, children could either learn about what colors, apparel, languages, or music they liked or communicate their attitudes or thoughts to the dolls. Consequently, the AIBD also learned and collected personal data, such as players’ personalities, emotions, attitudes, or behaviors. Finally, the data were stored in a private cloud repository, where only authorized parents could access the reports. In our findings, the average correlation between the capability of the AIBD’s learning performance and children’s active learning processes was high (support value equal to 0.791 and confidence value equal to 0.853). Moreover, the children playing with the AIBD could not only develop their emotions but they also saw large improvements in their ideas generation.