"For several days, Modern Aerobics has been called a lifestyle project of rock music with computer and realistic humans. We guarantee results – and guarantee them every time. This, which has also been referred to as Mod Aero (or ModAero), is known to revolutionize on occasion. At this unique workshop with team building and innovation in rhythm and movement, you will learn about the very foundation of a modern lifestyle and get the opportunity to achieve Perfect Intelligence (Deluxe Edition). Don't miss out. You have been selected to experience this truly unforgettable rock and computer music in a lifelike setting."

Modern Aerobics

Video loop generated with the Video Input function of Stable Diffusion based on stock footage of animated infographics

For the release concert of the genre-bending avant-pop band Modern Aerobics' debut album Perfect Intelligence, I created a set of conceptual live visuals, based on loops generated in the open-source deep learning, text-to-image AI model, Stable Diffusion. The loops were generated using a variety of techniques, using text prompts as well as image and video inputs. The loops were generated using various techniques, text prompts, as well as image and video inputs. The project is a distorted mashup of the visual language of contemporary online advertising trends, the Corporate Memphis art style, LinkedIn posts, and infographics devoid of information. 

I used AI-generated imagery to strip the source material from any coherent meaning, only leaving hints of attempted communication. By exploiting the flaws of the Stable Diffusion model - such as its (current) propensity to plunge any and all text straight into the uncanny valley - I could disjoin form from substance, leaving only a hollow shell of a corporate vernacular aimed to promote,  persuade and impress.
The "mood board" for the event, used also for costumes and choreography, was created in collaboration with Modern Aerobics, by feeding the full titles of advertising stock videos from Storyblocks.com as prompts into DALL·E 2

Video loop generated with the text prompt interpolation function of the Stable Diffusion, using AI-generated text prompts
with Corporate Gibberish Generator™ by Andrew Davidson

Video loop generated with the Video Input function of Stable Diffusion based on stop motion stock footage

Video loop generated with the Video Input function of Stable Diffusion based on stock footage montage

Video loop generated with the Video Input function of Stable Diffusion based on stock footage montage

Back to Top