Reframing biased AI-generated imagery: children as democratic co-interpreters in early childhood education

Artificial intelligence is entering early childhood education faster than theory has caught up. Recent scholarship has focused mainly on teachers’ perceptions, ethical risks, and implementation questions, while young children remain positioned predominantly as recipients of others’ decisions. This conceptual article argues that early childhood education should move from a deficit framing of children as passive viewers of potentially biased AI-generated imagery toward a democratic framing of children as capable co-interpreters of visual representation. Drawing on children's rights traditions, international guidance on children's rights in digital environments, a well-established model of meaningful child participation, the European Early Childhood Education Research Association's revised ethical code, and participatory early childhood scholarship, the article advances three propositions: that AI-generated imagery is pedagogically consequential because visual bias shapes children's sense of belonging; that young children can and should be invited into guided interpretation of such imagery; and that doing so constitutes democratic practice, not merely digital literacy. The paper concludes with implications for curriculum, teacher education, and policy, and proposes a research agenda for child-centred inquiry into AI in early childhood settings.