My Grandad used to write me letters every now and then, usually in pencil. He wrote about how he was doing; about his fishing or a car he was working on (he was a mechanic.) As the years went by, the letters got shorter. The last one he sent was upbeat but only a few sentences long. The handwriting was unsteady and the lines as faint as a spider web. His words said one thing but his handwriting said another. It occurred to me that he could have drawn a single line across the page and I would have had a sense for how he was doing.
Brushstrokes contain information separate from that which they describe. We instinctively read their weight and speed just as we read the words or forms they denote. We sense the movements that created them, and can’t help but associate those movements with a feeling or intent. A heavy, quick mark might translate into bold and confident. A delicate, slow one might translate into tentative and unsure. Each stroke of the brush represents a specific moment in an artist's thinking, a particular decision that conveys their response to both their internal and external environment.
In “What Painting Is,” James Elkins describes painting as “liquid thought” to illustrate how a paint stroke encapsulates movement and intent. This got me to thinking about AI-emulations of painting: if an AI brushstroke is neither liquid nor thought, what is it then? What does it represent or express? Since AI can’t think, and the human prompter has no control over individual brushstrokes, then an AI-generated brushstroke, by itself, is a simulacrum of thoughtfulness. Does this mean that it's devoid of any meaning; an empty gesture?
When standing in front of a human-made painting on a wall, we stand where the artist once stood in relation to the canvas when it was painted. It can feel as though we are looking through the artist’s eyes. It's intimate. Whose eyes are we looking through when we view an AI-emulated painting? Is it a combination of the human “prompt engineer” and the collective consciousness of all artists whose work serve as the foundation for the image? There is no easy answer to this and I suspect this is why AI emulations of paintings make us go a little cold. The AI image may be a deft imitation of human painting but the second we realize the underlying armature of brushstrokes is just the illusion of thought, the artist disappears and the meaning collapses.
Every new medium that comes along initially goes through an awkward phase of experimentation. Since early adopters don't yet know what to do with it, they naturally gravitate toward the aesthetic of existing mediums (thus today’s infatuation with AI-emulations of paintings.) For instance, some early photographers, known as Pictorialists, sought to imbue their photos with a painterliness by printing their photos on textured papers and posing their models in classic genre scenes. I think early adopters are also just testing how people react to the new medium. One of the first motion pictures was "Arrival of a Train at La Ciotat” which was first shown in a theatre in 1896. The silent short film is 50-seconds long and depicts a train pulling into a station. Apparently, the audience was thrilled and there are accounts of people running to the back of the theatre to avoid the oncoming train. Was it great cinema? No, but it wasn't meant to be. It was a demo. I think today we are like that theatre audience as we watch the latest AI imagery-- the work may not speak to us on an emotional level yet, but it is fascinating and thrilling to watch nonetheless. Eventually, the fog of novelty will burn off AI, and it’s innate qualities as a medium and tool will gradually emerge. As Susan Sontag wrote, "The painter constructs, the photographer discloses." What will the AI artist do?
If you haven't already signed up to receive my blog posts, you can do so here.