Abstract
Generative AI has proven itself as an efficient innovation in many fields including writing and even analyzing data. For spatial computing, it provides a potential solution for solving such issues related to data manipulation and analysis within the spatial computing domain. This paper aims to discuss the probabilities of applying generative AI to graph-based spatial computing; to describe new approaches in detail; to shed light on their use cases; and to demonstrate the value that they add. This technique thus incorporates graph theory, generative models to model spatial relations, generate new spatial forms and improve on spatial decision-making processes. The paper surveys such methods, describes typical applications, and outlines further development of the subject.