Matsys

Great review in by Allison Arieff on the exhibition “Transductions: Artificial Intelligence in Architectural Experimentation”. One of my images for the exhibition from the Urban Resolution series is featured in the review. In this series I use a very simple prompt “A street in X” with X being iteratively replaced with one of the 81 cities the UN has determined have a population of 5 million or more. Exploiting a bug in the Stable Diffusion model, I force the model to generate images far outside the resolution on which it has been trained. Although this process is normally avoided by users looking to produce coherent images, I’m drawn to these city streets in how they duplicate form and reveal the model biases. A Street in Tokyo becomes a wall of advertising and wires, while A Street in New York is yellow cabs and urban canyons. Seen here are the cities Tokyo, New York, Cairo, and Rio de Janeiro.