From bias to agency: empowering early childhood educators to challenge AI-generated stereotypes

Abstract Artificial Intelligence image generators are increasingly used to create educational resources, yet often inherit and amplify societal biases. In early childhood education, a field already shaped by gendered and cultural stereotypes, unchecked use of AI-generated images risks reinforcing narrow narratives about who educators are. This carries a dual risk: misrepresenting who educators are, and shaping the cognitive schemas children develop during critical periods of identity formation. Drawing on Freire’s critical pedagogy and Butler’s performativity of gender, we position educators as critically aware decision-makers rather than uncritical users of AI tools. Through conceptual synthesis of literature on AI bias, image generation, ECE stereotypes, and critical digital literacy, we develop a six-step decision-making framework guiding educators in recognizing, analysing, and challenging biases in AI-generated visual content. The framework addresses identifying bias in seemingly neutral technology, evaluating images for inclusive practice, and integrating critical evaluation into professional workflows. We outline strategies for professional practice and initial teacher education, ensuring emergent technologies serve as instruments of inclusion and agency rather than vehicles of stereotype and bias.