Tue.21.Apr.2026. Author: Thomas De Meester
Cities across Europe are eager to leverage AI to improve mobility, sustainability and public services. Yet in practice, many small and mid‑sized cities face the same dilemma: the models exist, but the conditions to use them do not.
Limited budgets, sparse sensor networks and fragmented data make it difficult for this group to translate promising AI research and applications into operational urban solutions. Meanwhile, much of today’s high‑quality AI research is developed and validated in and for large, data‑rich cities. This disconnect between research potential and real‑world applicability risks widening the AI gap between cities.
sing data and technology.
This whitepaper is the result of a series of experiments to address that gap. Its goal is to valorize existing research on transfer learning by testing whether AI models trained in data‑rich urban environments can be meaningfully adapted to cities with far fewer resources. Rather than asking smaller cities to start from scratch, the work explores how they can build on what already exists.
Using traffic prediction as a concrete and widely relevant smart city use case, the study applies this question in a real‑world setting. The whitepaper documents what works, what doesn’t, and under which conditions transfer learning can become a practical pathway for more inclusive AI adoption.
Thomas De Meester
E-mail: Thomas.DeMeester@imec.be
Co-funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them.