Storypair: Supporting Co-Reading in Bilingual Immigrant Families through Generative Language

Abstract

Reading is an essential yet challenging skill in child education, particularly for immigrant families navigating bilingual environments. This study explores how generative language models (LLMs) can enhance parent-child co-reading within these families. Informed by a formative study with ten immigrant families, I present the design, prototyping, and evaluation of StoryPAIR (Parent AI Integrated Reading), an interactive reading interface designed to complement physical books, for bilingual and immigrant families. StoryPAIR leverages language support widgets, LLM-powered question generation and answering, and learning session reviews to augment the co-reading experience. An initial user study with eight parent-child pairs from immigrant families indicates that StoryPAIR successfully engages both parents and children and enhances interaction flow.