Something broke in 2022 and everyone in the art world felt it, even the ones who pretended not to.
A door slammed open. Midjourney, DALL-E, Stable Diffusion—suddenly anyone with a keyboard and a vague aesthetic sensibility could produce an image that looked, by every casual measure, like it took skill to make. The forums filled up with breathless posts. The artists filled up with dread. And the commentators, as commentators do, filled up the space between with noise.
Most of what followed was the wrong conversation.
The wrong conversation is about whether AI art is “real” art. That question is a philosophical cul-de-sac — interesting enough at a dinner party, useless everywhere else. The right conversation is harder and more uncomfortable: what has AI exposed about what we were pretending art was, and what it actually is?
Here is the uncomfortable answer. A significant portion of what gets made, sold, and celebrated as visual art is surface. Competent surface, sometimes beautiful surface—but surface nonetheless.
Style without interiority.
Technique without necessity.
Work that looks like something was at stake, when nothing was.
AI can make that. AI can make it faster and cheaper and in greater volume than any human ever could. If that’s what your practice is built on, the problem isn’t AI. The problem was always there.
But a painting that could only have been made by one specific human being, at one specific point in their life — a photograph where the decision to be in a place, at a specific moment, to wait for that light, to see what most people walked past—that accumulation of experience, loss, obsession, and unresolved questions that no model trained on the entire history of human image-making can touch. Not because the machine lacks technical capacity. Because the machine lacks a life.
This is the distinction that matters. Not skill. Not style. Not even beauty.
That intent, the anxiety, the joy of pain—that necessity to express!
The work that survives this moment—and the work worth collecting right now—is work where you can feel that it had to be made. Where the alternative to making it was something the artist couldn’t live with. Where the image is not a demonstration of capability but a consequence of being human in a particular way.
AI produces images.
Artists produce existential evidence.
The collectors who understand this are not retreating from the market—they are moving through it with more precision than ever. Provenance has always mattered for authentication. Now it matters for meaning. The question is no longer only who made this. It is why this could not have been made any other way.
For us, AI is a condition, not a crisis. The same way photography was. The same way digital reproduction was.
Every one of those moments was supposed to end one form of art.
Every one of them instead burned away the reasons to execute badly, and left the real reasons standing.
This one will too.