Telephone Paintings
GAN, Procedurally Generated. Collaboration with Steven Sych.

Featured in Towards Data Science.

Telephone Paintings is a real-time generative video work that utilizes a neural network to create an infinite set of abstract paintings as well as a continually generated digital gallery to exhibit those paintings.The work deploys a generative adversarial network (GAN) which operates by (on the one hand) generating an enormous number of images and (on the other hand) criticizing its own productions on the basis of a set of provided ‘real’ works. The results are an infinitely varied set of paintings in the style of the works we’ve chosen to show the GAN; here, this means an infinite set of paintings in the style of 20th century constructivism-- an art movement whose prime concern was with the industrialization of artistic practices.

To curate and display these works, we have created a continually generating digital gallery environment. It is an industrial, corridor-like space, populated at random intervals with everyday objects: chairs, pillars, radiators, martini glasses. These objects pull at our attention, which is ever desirous of novelty. The digital gallery features an infinitely long wall upon which the AI’s paintings are exhibited. As the digital camera (i.e. the audience’s perspective) tracks along the corridor, the generated paintings appear in view. 

“I was not afraid of losing the “personal touch” so highly valued in previous painting. On the contrary, I even gave up signing my paintings. I put numbers and letters with the necessary data on the back of the canvas, as if they were cars, airplanes, or other industrial products.”

-László Moholy-Nagy (1944)