I Make Art With AI. Here Is What That Actually Means.

When I tell people that the NOA collection is made with Midjourney, I get one of two reactions. The first is genuine curiosity about how it works, what the process looks like, whether it is really art. The second is a kind of quiet suspicion that I am selling something I did not really make.

I understand both reactions. And I think both of them deserve a proper answer rather than a defensive one.

What the process actually looks like

Midjourney is a text-to-image model. You describe something and it generates a visual. In the most basic description, that is all it does.

In practice, using it to make work I would put my name on is considerably more involved than that description suggests.

I am not looking for a beautiful image in a general sense. I am looking for a specific quality of light. A particular way that illumination holds still in an image, or dissolves at its edges, or concentrates at its centre without becoming theatrical. That quality is what the entire NOA collection is built around, and finding it through a generative process requires a kind of sustained, iterative attention that I did not expect when I started.

A typical piece goes through somewhere between thirty and sixty iterations before I find something worth keeping. Most of those iterations are failures. Some of them are interesting in ways that redirect the search. Very occasionally one of them is exactly what I was looking for, and I know it immediately in the same way I imagine a photographer knows when they have the frame.

The decisions about what to pursue, when to stop, when something is finished are mine. The aesthetic sensibility behind those decisions is mine. The tool is Midjourney, in the same way that a camera is a tool or a darkroom is a tool. The work is not.

The question people are actually asking

When someone asks whether AI-generated art is real art, they are usually asking something more specific. They want to know whether there is a human behind it who actually cared. Whether the thing they are looking at is the result of a genuine aesthetic decision or a fortunate accident that required no judgment to arrive at.

Those are reasonable things to want to know. And I think they deserve honest answers.

There is a human behind NOA. I care about this work a great deal, which is obvious from the amount of time I spend on iterations that will never see the outside of my hard drive. The images are not random or accidental. They are the result of a specific, directed search for a specific quality of light, and the vast majority of what that search produces gets discarded.

Whether that constitutes authorship in the traditional sense is a genuinely interesting question, and I do not think it has a clean answer yet. The Guardian's ongoing coverage of AI in contemporary art reflects how unsettled the conversation still is at an institutional level. I find that more honest than sources that have already decided.

What I can say with confidence is that the work reflects a specific sensibility. You could look at the NOA collection and understand something true about what I find beautiful and why. That feels like the meaningful test.

Why the edition limit is not a marketing tactic

One of the real anxieties around AI-generated work is reproduction. The file exists. Printing it again is trivial. So what does it mean for something to be limited?

For NOA, the limit is thirty editions per piece. After the thirtieth print, the file is retired. Not paused, not held in reserve for a future reissue. Retired.

This is a commitment I made at the start and take seriously because I think it changes the nature of the object. When something genuinely cannot be made again, the relationship between the buyer and the work shifts. It is no longer just a beautiful image. It is a specific, finite thing. It is one of the things that separates a considered purchase from buying a poster, and that difference should mean something.

I am also aware that saying a work is limited and then finding reasons to print more of it is a betrayal of everyone who bought earlier editions on the basis of that promise. I am not interested in doing that. The limit is real.

The photography comparison I keep returning to

When photography emerged as a medium, it was not considered art. The camera did the work that the hand had previously done, and that made people suspicious of the result. The argument was about the tool, and it went on for decades.

Now nobody serious argues that photography is not an art form. The question of whether human vision and judgment and intention can produce something worth looking at through a mechanical instrument has been answered so thoroughly that we have forgotten it was ever a question.

I think we are somewhere in the early part of that same process with generative image-making. The tool is new and the questions it raises about authorship and labour are real and worth taking seriously. But the fundamental question about whether someone looked at this, cared about it, made decisions about it has not changed. And for the work I make, the answer is yes.

What the collection is actually about

Beyond the process question, there is something I want to be clear about regarding what NOA is trying to do.

The collection is not a catalogue of beautiful images. It is a coherent body of work built around a single sustained obsession with light and how it behaves in an image. Every piece that makes it into the collection shares that quality. Every piece that does not make it, regardless of how striking it might be on its own, gets left out because it does not belong to that specific conversation. The care that goes into the production side, what the matte acrylic surface does to an image and how it behaves in the shifting light of a real room, is an extension of the same commitment to coherence.

The result is a collection that is personal in a very literal sense. It reflects my taste, my eye, my particular way of seeing. And what I have found, perhaps unsurprisingly, is that the people who respond to it most strongly are the people who share something of that sensibility. Who have the same quiet preference for restraint over spectacle, for light that holds still rather than performs, for something on their wall that asks them to look rather than demanding to be looked at.

That is who I make this for. Not everyone, and not trying to be. If you have read this far and something in what I have described feels familiar, the collection is probably worth your time.

NOA is a Paris-based limited edition art studio. Every piece is made through an iterative process that I direct, limited to 30 editions, and printed on museum-grade matte acrylic. When an edition closes, it closes.