Generative Artificial Intelligence in the Context of Urban Spaces

Urban areas face various challenges, including urban planning, waste management, pollution, public safety, and other issues. Information technologies and artificial intelligence have proven helpful in addressing these problems. This paper explores the potential of Generative Artificial Intelligence (GenAI) to address urban challenges through geospatial analysis. We cover various GenAI models used in urban contexts, such as Generative Adversarial Networks (GAN), Variational Autoencoders (VAEs), Transformer-based models (or General Pre-trained Transformer), Large Language Models (LLMs), and Generative Diffusion models. The paper explains how GenAI can be used for urban analysis applications from an information management perspective. It presents examples of the operations that can be achieved. Additionally, it describes studies related to energy and resource management, urban planning, natural disaster management, and traffic management, demonstrating the advantages of applying this type of technology. The use of GenAI, along with Geographic Information Science and Technology (GIS&T) and Smart Cities concepts and tools, is being discussed. It presents some points to consider when designing, developing, and implementing GenAI-based applications for urban space analysis.