In machine learning, an ablation study is a way of examining what happens when parts of a system are removed, eroded, or destroyed by their human experimenters. Ablation Studies turns this logic on its head, examining how systems of human meaning and relations degrade under the proliferation of artificial intelligence, how their subjects push back, and what new beauty might emerge from the dialectic.

Here Esherick turns his signature “latent painting” process toward works from the canon of art history to reflect on the impact of artificial intelligence on art and image-making culture. The method disrupts the image generation process to look inside the “mind” of the AI model, pulling out half-formed images that we were never intended to see. Where the artist previously used the technique to create entirely new images, here existing images are fed to the model and prematurely removed, using this glimpse inside the system to explore what it means for images to be consumed, digested, absorbed, and spit back out by it. The aesthetic of erasure that’s conjured beckons us to observe the history of human culture through this lens, as a kind of psychic landscape populated by beautiful, haunting ruins.

The resulting series embraces deeply conflicting emotions toward the changes wrought by technology. On the one hand, we see in these images the history of art being eroded, obscured, and co-opted before our very eyes, the hands of its artists all but erased. And yet the new aesthetic of these abstractions offers a sense of the poetic to accompany the loss, and a sense of hope that these new systems will complement rather than replace what they ingest. In change, there is growth. In rupture, we find rapture. In death, rebirth.