AI Sandbox
One of the most challenging and exciting aspects of an architectural design process is the ideation phase, the period of experimentation with form with only a few global restrictions. This is also a critical phase for the so-called “big moves” of the project. Before the digital design era, this was done through the “blue foam approach,” cutting a foam modeling material to produce as many variations of the form as possible. Today, software operations have replaced the material procedures of blue foam with mouse and keyboard interaction patterns, and the foam approach is replaced by simulation and procedural modeling techniques.
While digital design and the advent of proceduralism have enabled increased variation during the ideation phase, there remain significant inefficiencies and labor-intensive processes, in particular, with regard to collaboration among multiple designers in a firm and resolution of project phases. AI Sandbox takes the proceduralist “blue foam” approach to form making into the realm of AI to amplifying and extend the ideation phase and automate the development phase. The model learns the tendencies and style of designers over time and can simultaneously propose insight into multiple resolved models. We are proposing a new kind of AI-assisted, supercharged, predictive design process for architecture that builds on existing procedural workflows to deliver the next modeling paradigm. This project develops software that streamlines the design process by leveraging machine learning AI to learn and replicate user patterns, focusing on user interaction and personalized content creation. The software will augment a designer’s initial formal decisions by layering predictive representations of project resolution in real-time, amplifying a designer's inputs. The software interprets the initial inputs and automatically generates things like plans, facade studies, and programmatic diagrams. This allows the architect to understand and explore the larger consequences of initial design moves. The goal is to speed up and automate the collaborative design process while allowing for the explosion of AI-enabled creativity that maintains the validity of human design ideation.
Funded by: Epic Games