This paper introduces the concept of Generative Digital Twins (GDTs), an evolution of Digital Twins (DTs) that incorporates Generative AI to enhance prediction, control, optimization, and simulation capabilities. Originating in the manufacturing sector, DTs have been revitalized by advancements in IoT and AI, allowing them to interact with real-world data and achieve their goals more effectively. The paper defines GDTs, explores how Generative AI bridges model- and data-driven approaches, and highlights its benefits in IoT environments, particularly within Smart City scenarios where predictive accuracy, system robustness, and explainability are crucial.

The paper discusses the importance of Digital Twins (DTs) enhanced by AI, Edge Computing, and IoT. It introduces the concept of “opportunistic” interpretation of DTs, creating dynamic digital replicas of physical objects through AI at the network edge. This approach is demonstrated through a traffic prediction use case, highlighting improved performance and resource efficiency.