Her name is Diane. I'm not using her last name because she asked me not to, and honestly, after what she told me, I understand why.
Diane runs a mid-size contemporary gallery on the north side... not one of the flashy River North spots built for Instagram foot traffic, but the kind of place that's been showing emerging Chicago artists for fifteen years. The kind of place where the walls have real opinions. She's not a tech person. She still keeps show notes in a spiral notebook. Her gallery cat is named after a dead sculptor.
Last spring, over coffee at Hopewell Brewing in Logan Square, she told me something she hadn't said out loud to anyone in her industry.
She'd been using AI to draft artist statements for three consecutive shows... and nobody had noticed.
The Setup
Diane has been in the gallery world since the mid-2000s. She's worked with artists across every stage of their careers, from first shows in Pilsen basements to representation at Art Basel. She knows what makes a strong statement. She also knows what makes artists freeze when they have to write one.
The truth is, most visual artists are not writers. That's not a criticism. It's just the reality. The people who are extraordinary with light, with texture, with the physical presence of a material... they are often the same people who stare at a blank document for two weeks and then email you something at 11pm that reads like a college application essay.
Diane's gallery runs four to six shows a year. Each show needs artist statements, wall text, press materials, grant application language, collector emails, social captions. For a one-person operation with a part-time assistant, that volume of writing is genuinely punishing.
Nobody tells you this about running a small gallery. The writing never stops.
The Problem
Last February, she had two shows opening six weeks apart. One artist handed her a statement that was technically accurate and completely lifeless. The other artist... didn't hand her anything. Just kept saying it was coming.
She had a press deadline. She had a printer deadline. She had a collector preview in ten days.
She opened ChatGPT, pasted in the first artist's bio, three images of the work, and a voice memo she'd recorded after walking the studio. She typed something like: help me draft a 200-word artist statement that sounds like this person made it, not like marketing copy.
It took four iterations. She rewrote about 40 percent of the final version herself. But the bones were there in twenty minutes instead of two weeks.
She did the same for the second artist. Then the third show. By the time she told me about it, she'd done it five times across three exhibitions.
The artists read the drafts and said yes. A few made small edits. One said it was the best statement she'd ever had. None of them asked how it was written.
What She Did
I want to be precise here because the framing matters.
Diane was not generating AI art. She was not replacing artists or their ideas. She was doing what every good editor, publicist, and gallery director has always done... translating the work into language. The difference is she had a drafting partner that never sleeps and doesn't charge $150 an hour.
Her process, as she described it:
First, she'd spend real time with the work. Studio visits, or if that wasn't possible, a long call. She'd take voice memos of her own reactions... what she saw, what she felt, what it reminded her of that she couldn't quite name. She'd gather any existing materials from the artist. Then she'd feed all of it into the AI with a very specific prompt: write in the first person as the artist, avoid jargon, avoid anything that sounds like a grant application, focus on the physical reality of the materials.
Then she'd edit what came back. Hard. Cut the sentences that felt performed. Add the specific detail only she knew from being in the room. Rewrite the opening every single time.
She wasn't lazy about it. She was efficient about it. There's a difference.
The Artsy survey of more than 300 gallery professionals found that a significant and growing number are using AI tools for exactly this kind of work... writing, research, donor communications, collection management. The gap between what the art world says publicly and what it does privately is wider than most people want to admit.
Diane is not an outlier. She's just the one who said it to me out loud.
What Happened
Three things happened that she didn't expect.
First, her press coverage improved. Not dramatically. But two publications that had ignored her previous shows ran small previews. She thinks the tighter, cleaner statement language made the publicist's job easier. The work was the same quality it had always been. The words around it got sharper.
Second, her artists started trusting her more with their language. Because she was producing cleaner first drafts, they had something concrete to react to instead of a blank page. One photographer she works with told her the process had finally made him understand what his own work was about. That hit different, she said.
Third... she started feeling guilty about it. Not immediately. But over time. Sitting with it.
She wasn't sure if she owed the artists disclosure. She wasn't sure if she was somehow diminishing the authorship of their statements. She'd spent fifteen years thinking of herself as someone who helps artists find their voice. Was she outsourcing that now? Was the voice still theirs if the first draft came from a machine she was prompting?
She doesn't have a clean answer. I don't either.
What I Learned
I've thought about this conversation probably thirty times since that afternoon in Logan Square. Here's where I landed.
1. The writing around art has always been collaborative... we just didn't call it that.
Gallerists, publicists, curators, editors... they have always shaped how artists are described to the world. The artist statement as a form is already a performance of a self, not the self. AI didn't invent the gap between who an artist is and how their work gets framed. It just made the drafting faster.
2. The Artsy survey isn't the scandal. The silence around it is.
More than 300 gallery professionals. Real numbers. Significant adoption. And the art world's public conversation hasn't touched it in any meaningful way. That silence is its own kind of answer. The institutions that police authenticity the hardest are the same ones using these tools the quietest. That's worth sitting with if you're an artist trying to figure out where you stand.
3. The question isn't whether you use AI. It's whether the final thing is true.
Diane's statements were true. They were accurate to the work, grounded in real observation, edited by someone with fifteen years of reading and writing about visual art. The AI was a drafting tool, not a replacement for judgment. That distinction matters more than the tool itself.
I shoot with a Fuji X-T5. I've used Lightroom presets, AI masking, noise reduction algorithms. Nobody asks me if my photographs are real because the light I stood in was real. The edit is part of the craft. So is knowing when to stop editing.
4. Artists deserve to know how their public language gets made.
This is where I think Diane's guilt was pointing somewhere useful. Not because she did something wrong, but because transparency is part of the relationship. Not every artist would care. Some would be relieved. But they should have the choice. The conversation about AI in creative work is still happening mostly between gatekeepers, mostly quietly. That's backwards. The artists whose work is being described, represented, sold... they should be in that room.
The truth is, the gallery world has always had a language problem. Statements that sound the same. Wall text nobody reads. Press releases written to impress other curators instead of actual humans. If AI helps fix some of that... I don't think the problem is the tool.
The problem is who gets to decide how we talk about it.
Diane is still using it. Still editing heavily. Still spending real hours in studios. She told me last month she finally disclosed to one of her newer artists and the artist shrugged and said, honestly I was going to ask you if I could do the same thing for my grant applications.
Which is about where we actually are.