Julian van Dieken is clear about what he is, and it's not an AI artist. He's a photographer and educator with 15 years of professional experience doing documentary work for non-fiction publishing, co-founding an educational training company, and teaching workshops on video production and project documentation.
In September 2022, he started a side project on Instagram — @julian_ai_art — with no particular goal. He describes it on his site as "a side project where I experiment with AI tools and try to learn something new."
That's literally the whole brief. Experiment. Learn. Share.
A few months later, his AI-generated image was hanging in the Mauritshuis Museum in The Hague, where Vermeer's Girl With a Pearl Earring normally lives.
What happened at the Mauritshuis
The museum loaned out the original Vermeer and ran an open competition — #MyGirlWithAPearl — inviting anyone to submit their own interpretation. Three thousand five hundred people submitted. Five prints were selected and displayed in the gallery. Van Dieken's A Girl With Glowing Earrings, created with Midjourney, was one of them.
He called it "crazy" and "completely surreal." He said: "One of the most famous paintings in history is literally being replaced by one of my AI images."
The internet was less charmed. People were angry that an AI-generated image had made it into a traditional fine art institution.
Van Dieken had been transparent about his method from the submission itself — he wrote in his entry that he was reflecting on how AI tools might change creative processes. The museum knew what it selected. He wasn't hiding anything.
The 30-Day Challenge as a method
The actual practice that produced that image is worth paying attention to.
His core method is the 30-Day Challenge: make something every day for a month, strip out professional pressure, and learn through mistakes rather than through caution. He uses it to force daily creation that builds real fluency with a new tool.
It works because you exhaust the ideas you were comfortable with and start reaching for ones you weren't sure about. That's where things actually happen. I recognize this from other creative practices — the ratio of attempts you'd share to attempts you make is part of what builds your instincts, not a side effect you hide.
He works primarily with Midjourney and Photoshop for post-processing, and has experimented with Stable Diffusion, Astria.ai, and others over time. His photography background shapes what he's trying to pull out of these tools — the Mauritshuis piece was described as "sharp, photorealistic," which makes sense for someone whose creative instincts formed around camera work.
He also says this plainly on his site: "I'm not a developer, I'm not a trained AI expert." The point of sharing his work and learnings is that he came to this from the creative side, not the technical side. He figured it out the same way a working photographer or designer would have to.
On the backlash
You can make a real argument that AI-generated work in a fine art museum is a category error — that it conflates the outputs of a generative model with what it means to make something, and that institutions presenting them side by side muddy distinctions that actually matter. That's worth taking seriously.
But van Dieken wasn't slipping anything past anyone. He disclosed the method in his submission. He reflected on it publicly. The museum ran an open competition with no stated restrictions on technique, and five prints were chosen from 3,500. If the problem is with the criteria, that's the museum's call to make and answer for.
The framing of him as someone who "ripped off" Vermeer using AI doesn't hold up against what actually happened. He submitted to an open competition, was selected, and was honest about how the work was made the whole way through.
What he's actually doing now
Past the Mauritshuis moment, the educator piece has become the main thing. His LinkedIn describes him as "AI Educator & Speaker." His work has been covered by the New York Times, Zeit, CNN, and Der Spiegel. He speaks at events — including MIS ArtFest 2024 on the topic of how to be creative with AI. He runs workshops.
He's not primarily someone who makes AI images. He's someone who made AI images specifically to understand the tools well enough to teach other people how to think about them.
That's a different function than "AI artist," and right now it might be the more useful one. There are plenty of people generating images. There's less of people who came to these tools from a creative background, put in the daily reps, and can explain what the process actually involves to someone who hasn't done it yet.
He set out to learn something new. What he ended up with was a museum wall, a following, and a set of conversations he now gets paid to lead. That's a longer answer to the original brief than he probably expected.