The first AI-generated image won an art

Jason Allen submitted an image called “Theatre D’opera Spatial” to the Colorado State Fair fine arts competition. It won first place in the digital arts category. When it came out that the image was generated using Midjourney, the internet lost its mind.

Artists were furious. “This is the death of art,” people said. “AI is stealing from human creators.” The discourse was loud and predictable and, I think, focused on the wrong thing.

The interesting part isn’t that AI won an art competition.

The interesting part is that the judges couldn’t tell.

The Turing test at a county fair

Three judges evaluated the submissions. They scored on technique, composition, and impact. Allen’s piece scored highest. None of the judges knew it was AI-generated.

This is, in a practical sense, a Turing test for visual creativity. Can a machine produce an image that human experts, evaluating on artistic merit, can’t distinguish from human-created art? At the Colorado State Fair, the answer was yes.

The test wasn’t designed. Nobody set out to run it. Allen submitted the piece because he thought it was good. The judges agreed. The reveal came after.

There’s something wonderfully accidental about that. The Turing test for art didn’t happen in a lab with controlled conditions. It happened at a state fair, between entries of oil paintings and digital illustrations, judged by people who just liked the picture.

The outrage

The New York Times covered the backlash. Artists on ArtStation posted protest images. The argument was that AI art is theft because the models were trained on human art without consent or compensation.

That argument has merit. The training data question is legitimate and unresolved. If Midjourney learned to paint by studying millions of human paintings, and those painters didn’t consent to being training data, that’s a real ethical and legal issue.

But the outrage extended beyond the training data question into something more existential. The feeling was: if a machine can win an art competition, what’s the point of being an artist?

I think the answer is the same as it was when photography appeared: the point of being an artist is whatever it was before, minus the parts that a machine can do. The parts a machine can’t do are the ones that matter.

What a machine can’t do

Allen spent 80 hours working with Midjourney to create the winning image. He ran hundreds of generations. He refined prompts. He made artistic decisions about composition, color, mood. The tool generated the pixels. The human directed the vision.

Is that art? I think so. The same way a photographer who didn’t build the camera, grind the lens, or coat the film makes art through the act of seeing and choosing.

The question that matters isn’t “can AI make art?” It’s “what do we value about art?” If we value only the technical rendering, AI wins. If we value the human act of deciding what’s worth rendering, humans aren’t going anywhere.

The judges at the Colorado State Fair valued the image itself, not its origin story. And honestly? I find that kind of beautiful. A picture was good. It won. The method of creation was revealed later, and people had feelings about it. But the picture was still good.

The Turing test for creativity happened at a county fair and the machine passed. The rest is commentary.

Related thinking: