The rapid emergence of Generative Artificial Intelligence (Gen-AI) is transforming how students in engineering courses learn, design and create. In higher education, this shift raises a fundamental tension on how institutions can preserve mastery of disciplinary fundamentals while preparing graduates to collaborate effectively with intelligent systems. This paper aims to provide a conceptual pedagogical framework named SPARK within the CDIO Innovation Process that guides educators in redesigning teaching and assessment in which Gen-AI acts as a trusted partner in thinking, creating and learning for students. SPARK represents Scope the Problem, Propose Solutions, Assess Critically, Refine and Integrate, and Knowledge Artefact and Presentation. Each phase corresponds to a CDIO stage where students conceive by scoping authentic problems, design by ideating and proposing solutions, implement through testing and refinement, and operate by communicating artefacts and reflections. The model is grounded in constructivism, self-regulated learning, reflection-in-action and authentic assessment, ensuring that pedagogy rather than technology anchors Gen-AI integration. SPARK also advances several CDIO Standards. It strengthens Standard 2 (Learning Outcome) by embedding Gen-AI literacy and ethical reasoning. It can enhance Standards 7 and 8 through active learning and integrated learning experiences and it supports Standards 9 and 10 by providing a vehicle for faculty competence and pedagogical innovation. As a design-based reasoning framework, SPARK moves fluidly across contexts and remains resilient to technological change. Its focus on inquiry, reflection and ethics renders it relatively immune to rapid disruption. Future implementation at Singapore Polytechnic will test SPARK in authentic modules, examining its impact on student agency, self-directed learning and cognitive engagement. Findings will inform its effectiveness and lessons learnt. In doing so, this work contributes to CDIO’s global mission of preparing engineers who can conceive, design, implement and operate responsibly in a Gen-AI enabled world.
