AI systems are pattern-finding machines. Upon feeding one a set of images, visual pattern recognition is achieved in part through hierarchical extraction of features. If a set of data—in this case, paintings—has consistent features, it’s easier for the system to learn, and replicate, the pattern.

I began thinking of “Old Masters'” works as pattern and asked myself whether I could extract their consistent elements and accumulate them as a mass of so-called genius. I compiled existing AI-generated images of figures in the “style” of William-Adolphe Bouguereau. Once I had a large enough set, I picked out the body parts the AI most confidently replicated. I cross-referenced them with actual Bouguereau paintings in order to assess the AI’s accuracy to the originals and decided upon a small library of body parts: toes feet, shoulder (trapezius, sternocleidomastoid, and suprasternal notch), nostrils, and corner of the mouth.

For this specific exhibition, I made four small pencil drawings. Instead of a more traditional paper, I chose polypropylene (a type of plastic) and used the AI-generated images as direct, literal references. They are in contrast/tension with the abstractions, used in the same (deadened) spirit as generative AI more broadly.

In this work, I’m considering and connecting three main ideas: how the AI bubble, amongst other displays of violence, drives water scarcity and increases carbon emissions; how hyper-categorization and binaries are structures of white supremacy; and the myth of individualistic merit and achievement in art and art history.