New Perspectives in Science Education

Grounded in Vygotsky’s sociocultural theory, this research re-examines the construct of More Knowledgeable Others (MKO) in family-based science learning when both parents and Artificial Intelligence (AI) are serving as learning companions. The emerging landscape of Generative AI for Family Learning changes the way children acquire understanding, curiosity and emotional engagement in the informal science learning context. In a comparative case study, six children (aged 9–11) engaged with an inquiry question with the help of AI and parents separately. The study aimed to investigate how AI- and parent-supported learning influence children’s science understanding, curiosity, and learning experiences in science. We analysed children’s interactions with AI and parents through observation, interviews and children-AI chat histories. Results suggest that AI plays a role of cognitive MKO by providing comprehensive factual scientific explanations to children whereas parents struggled in such a role and tended to provide brief and sketchy answers, jokes or theological reasonings. However, children showed high trust to parents as affective MKOs and only gain about 47% scientific knowledge out of the excessive information provided by AI. These results suggest that AI and parents, with their distinct yet complementary strengths, can be integrated into a synergistic GenAI-assisted family learning model that balances cognitive scaffolding and emotional connection to enrich children’s science learning within sociocultural spaces.