Despite the abundance of urban data available and the possibility of improving decision-making city administration skills, and infrastructure, contemporary data-driven methods for city data-based knowledge discovery cannot frequently use collective data. In a loose manner specified elements for data interpretation, or disciplinary discrete analyses of particular datasets make it simple to ignore essential subject knowledge, frequently leading to speculative decisions. Intelligent City Digital twins are intended to get past this obstacle by incorporating analytics and visualization that are more comprehensive approaches to knowledge discovery in real-time procedure from diverse city-data. Here, we offer a spatiotemporal knowledge discovery framework that integrates social and sensor data and allows for insights from human cognition for the collective exploitation of city data in smart city digital twins. Even though these applications are effective, they frequently use vast quantities of high-dimensional and multi-domain data to monitor and characterize various urban sub-systems. This poses problems in application areas where data availability and quality are restricted, as well as expensive efforts to generate urban scenarios and design alternatives. Generative Artificial Intelligence (GenAI) models are a new field of study in deep learning that has shown promise in producing original material. With an emphasis on several urban sub-systems, including building and infrastructure, energy, water, and transportation, this article attempts to investigate the creative integration of GenAI approaches with urban digital twins to address issues in the planning and administration of built environments.