Towards Gamifying Interactive Language Learning using Large Language Models for Children

As Large Language Models (LLMs) rapidly evolve, they hold untapped potential to reshape language learning for children, transforming learning into interactive, tailored experiences. This study presents CELLA (Children’s Edutainment and Language Learning Assistant), an innovative conversational system that empowers young Chinese learners through LLM-powered gamification. CELLA empowers young learners to acquire and apply Chinese language knowledge and skills by interacting with LLM and creating customized educational gaming content. Our approach integrates traditional text-based language models with multimodal elements, including images, audio, and four specially designed educational games. Through a user study involving 52 children aged 9 to 10, we demonstrate that CELLA enhances engagement and learning outcomes compared to plain LLM-driven conversational interface. We believe that developing interactive, LLM-powered systems like CELLA contributes to the growing body of research on children-AI interaction, opening new avenues for AI-driven educational technology.