Objectives: This study qualitatively explores how young children develop an emerging understanding of AI literacy and ethical awareness through play-based interactions with AI robots. It examines how a multidimensional AI literacy construct, integrating affective, behavioral, cognitive, and ethical components, unfolds within authentic early childhood educational contexts.Methods: The participants included 103 children (aged 4-5) and four teachers from two kindergartens in D City, South Korea. Ten sessions of play-based activities, centered on understanding, applying, evaluating, and ethical awareness were conducted using an Alpha Mini robot (NAVER CLOVA). Data sources, including video recordings, observation notes, teacher journals, and focus group interviews, were analyzed through thematic analysis of interaction episodes using MAXQDA 2024.Results: Three main themes and nine sub-themes were identified. First, the children demonstrated an emerging understanding of AI mechanisms, recognizing technical limitations while perceiving robots as emotionally relatable entities. Second, they employed strategic behaviors, such as modifying speech and regulating turn-taking, to overcome technical challenges. Third, children exhibited ethical responsibility as active agents, valuing information accuracy and practicing relational ethics of care. Teachers played a crucial mediating role, scaffolding interactions and guiding reflection without directing outcomes.Conclusion: These findings indicate that young children can develop foundational AI literacy that integrates technical comprehension, communicative adaptation, and relational ethics. This study highlights the need for early AI education to embrace technical errors and limitations as learning opportunities, foster collaborative and reflective communication, and extend relational care into ethical responsibility. Such insights contribute to designing early childhood curricula that nurture active, reflective, and ethically aware digital citizens.
Bae et al. (Thu,) studied this question.