Public Space Design enabled by Artificial Intelligence Generative Tools

Abstract

This paper discusses the integration of generative artificial intelligence (AI) tools in the conceptual phases of the design of a public space. It raises the question of how AI tools can enhance the design process and how the integration of AI tools in a design process may affect the interactions between architects and clients. The investigation uses a collaborative project between academia and a major art institution to transform a vehicular street into a pedestrian event space supporting art and community programs.

The methodology was framed around generative AI platforms, such as Dall-E 2, Midjourney, BlueWillowAI, DreamStudio, Lexica, and Stable Diffusion utilizing both text-to-image and image-to-image functionalities. The approach followed the reiterative process of developing design variations, evaluating outputs, identifying desired traits in imagery/designs, and exploring new possible designs based on desirability. The study demonstrated (1) accelerated idea formations in the early stages of the design process and (2) an enhanced ability to communicate with a broad audience.