AI literacy for AI natives? A four-country comparison

Abstract Artificial intelligence (AI) is becoming increasingly mainstream in children's homes and everyday lives. Using neo-ecological theory as a conceptual framework and Yang et al.'s (2026) five-component AI literacy progression model as an analytical lens, this document analysis examined core ECEC curricula and related guiding documents currently in use in Australia, China, Japan, and South Korea to determine whether early childhood education and care (ECEC) guiding documents are keeping pace with the rapid development. The analysis reveals that the curricula of Japan and South Korea contain no explicit reference to AI literacy, prioritising human-to-human relationships and offline safety. China's national ECEC guidelines similarly contain no AI references. However, Shanghai's strategic action plan explicitly positions kindergartens within broader AI education initiatives, though without specifying pedagogical progressions for young learners. Australia demonstrates limited and uneven alignment across AI literacy components, with the strongest alignment in “Interacting with AI Agents” and no alignment in “Understanding Human-AI Relationships”. Young children’s development of AI literacy is thus likely to remain contingent on the home environment, where they may be interacting with AI agents and creating artefacts more than developing awareness of ethical considerations and risks. Wide differences in children's foundational AI literacy capabilities may entrench educational disadvantage along the lines of AI literacy as it becomes increasingly mainstream in society. We argue that curriculum-guiding documents should be strengthened by articulating explicit AI literacy learning priorities and developmental expectations for young children, accompanied by culturally responsive resources to equip educators to facilitate equitable learning.