Haridis Opens 5-Part MIT Exhibition Probing AI, Aesthetics and Architecture

1 articles · Updated · MIT News · Jun 29

Through June 30, MIT’s Keller Gallery is showing Alexandros Haridis’ “Beyond Data-Driven Aesthetics,” a research exhibition that turns theories, algorithms and machine-learning systems into physical installations and interactive visualizations.
Five thematic sections trace how computing has been used to judge and generate aesthetics in architecture and applied arts, from George Birkhoff’s 1930s mathematical measure of beauty to AICAN’s novelty-based image evaluation.
Haridis said the project grew from watching tools such as ChatGPT and Stable Diffusion reshape public debate around creativity, while highlighting that many of those questions were already present in AI’s 1956 Dartmouth origins and earlier design-computation research.
The exhibition argues design itself can serve as an interpretive method, using software reconstruction, fabrication and data visualization to make opaque computational systems more legible for broader work on the built environment and human experience.