I gave a talk at metaLAB (at) Harvard, in the Cambridge campus of Harvard University. The metaLAB is represented by a community of scholars, designers, artists, makers, technologists, curators, and educators and consists of three locations – Harvard University, the Freie Universität Berlin, and the Academy of Arts and Design Basel. The hosts of the talk were Sarah Newman (Principal at metaLAB), Austin Ledzian (Researcher, interim Director), and Jeffrey Schnapp (Founder at metaLAB).

Abstract

Along with the exponential increase of data production and the development of fast computing and graphical processors, the past few years have seen a substantial increase in effort toward the mechanization of creative processes in design, art, science, and technology. Quite a number of articles describe computer programs for game-playing, art creation, literature, design generation, visualization, and scientific discovery. Generative AI, the catchphrase underneath the hype, is a byproduct of engineering achievements in “deep” neural networks.

This talk focuses on the conceptual meaning of the term “generative” and the computational interpretations it receives in theories of design computing in architecture and related areas of spatial design. Understanding the term in this way, and some of its historical origins, will enable architects and researchers in the area of computing to illuminate various controversies that rage today – between cognitivist and connectionist approaches to creativity, between authorship and forgery, between “rules” and “data.”