Preparing Preservice STEM Teachers for AI-Driven Classrooms: Doing Teacher Education Programmes Differently Now

In this manuscript, we argue that preparing preservice STEM teachers for AI-driven classrooms requires urgent and systemic change in higher education. Artificial intelligence is rapidly transforming pedagogy by enabling adaptive, student-centred 
learning, but many teacher education programmes in Nigeria and beyond remain rooted in pre-AI paradigms. We propose a coherent framework that embeds AI literacy, cultivates high-quality situated practice opportunities, builds teacher educator capacity, enacts policy and institutional reforms, and anticipates ethical and professional risks. Drawing on the quadruple helix model, we emphasise the importance of sustained collaboration among educators, researchers, industry, and policymakers to co-create curricula, policies, and governance strategies. We highlight practical approaches, including scaffolded AI literacy courses, sandboxed microteaching, lecturer professional development, and policy-backed institutional structures that ensure sustainable adoption. At the same time, we address risks such as bias, inequity, privacy breaches, and teacher deskilling, advocating for layered mitigation strategies that protect teacher agency and learner equity. This framework positions preservice STEM teachers not just as technology users, but as reflective professionals capable of leveraging AI responsibly, critically, and creatively. By integrating innovation, ethics, and systemic alignment, we present a roadmap for future-proofing STEM teacher education in Nigeria and similar contexts.