Beyond inclusion: feminist AI for transformative justice in STEM

Abstract While human-centred AI (HCAI) has emerged as a commonly deployed response to algorithmic harms, it remains largely reactive and limited in addressing their structural causes. Drawing on feminist STS, intersectionality, critical posthumanism, and more-than-human anthropology, the paper proposes a more comprehensive Feminist AI Framework (FAIF) that expands existing approaches by addressing both immediate biases and the onto-epistemological assumptions sustaining them and positions feminist analysis as integral to technical literacy. Drawing on extensive classroom exercises using text-to-image generators, the paper argues for the systematic integration of feminist perspectives into the development and deployment of AI and in STEM education in general.