Search
Position statement · Version 1.0
We are publishing this before anyone here has asked us to. There is no local dispute we are settling—we would rather say plainly what we think now than improvise it later, under pressure.
CUBE exists to catalyze artists. Everything below follows from that and from nothing else.
We are a young foundation in a small city with a serious arts community, and we are publishing this before anyone here has asked us to. There is no local dispute we are settling. We would simply rather say plainly what we think now than improvise it later, under pressure, in the middle of an argument about a particular artist’s particular work.
01
The work we fund, exhibit, and put in front of this community must originate in a person with something to express. That is the standard, and it is the only one that has ever mattered. It was the standard when the tool was a chisel. It is the standard now.
02
This is not a concession we are making to technology. It is a basic condition of artistic freedom. An artist who wants to grind their own pigment should grind their own pigment. An artist who wants to use a model at three in the morning to find the composition they have been chasing for a month should do that. Neither owes anyone an apology for the choice.
03
We will not ban AI from the work CUBE funds, commissions, exhibits, or acquires. Our reasons follow, and they are not “we like technology.”
04
Tell people how the work was made—plainly, without ceremony and without apology.
05
CUBE’s values are Culture, Unity, Belonging, and Equity. A tool that lowers the barrier to making, researching, translating, and taking part is not incidental to those values. It is close to the center of them.
We would rather state the case against us properly than pretend it is weak. It is not weak.
The large image models were built by ingesting enormous quantities of work, including work by living artists who were never asked, credited, or paid. That is true. It is not answered by noting that the courts have yet to finish deciding whether it was lawful. The courts may well conclude that it was—and the artists whose portfolios were scraped would still be owed something they never received. That is a moral claim, and it survives whatever the courts decide.
This is not a hypothetical harm arriving at some future date. In a 2024 study commissioned by the Animation Guild and the Concept Art Association, roughly three-quarters of surveyed entertainment executives said generative AI had already contributed to eliminating, reducing, or consolidating roles in their own divisions. A UK Society of Authors survey found about a quarter of illustrators reporting work already lost. Concept art, storyboarding, and illustration are absorbing this first and hardest. Any arts organization that discusses AI without saying so out loud is not being serious.
The strongest objection is not about law or jobs. As the computer-graphics researcher Aaron Hertzmann has put it: every previous tool left an audience able to tell roughly how a thing was made. Nobody mistook an early photograph for a painting. Nobody mistook a synthesizer for a string section. Generative AI is the first tool where that signal disappears—the viewer can no longer read a work’s making off its surface, and a great deal of how we assign meaning, effort, and value depends on being able to.
We think this objection is correct. We simply draw a different conclusion from it than the people calling for a ban. If the problem is that provenance has become invisible, the remedy is to make provenance visible again. That is disclosure. It is not prohibition.
The genuine injury is upstream, in how the models were trained and who was not paid. Turning down an artist in Healdsburg returns nothing to a concept artist in Los Angeles whose portfolio was scraped in 2021. It alters no training set. It changes nothing about the thing people are rightly angry about. It only forecloses the artist standing in front of us.
These capabilities are now embedded in ordinary software—in Photoshop, in Lightroom, in the phone that took the reference photograph. There is no line an administrator can draw and then defend. What fills the vacuum is suspicion: juries squinting at hands, submissions litigated on the internet, artists accused on the evidence of an odd thumb. This has already happened elsewhere. It punishes the honest, rewards the quiet, and turns a community of practice into a community of inspectors.
Prohibition preserves an existing picture of what an artist is and does. It does not protect any particular artist’s ability to grow, pivot, or change their practice—it constrains it. An artist who decides their work now involves these tools has not stopped being an artist. A rule that says otherwise has stopped being about art and started being about status.
We understand the impulse. It comes from real loss and real anger, and the people who feel it are not fools—many of them are the artists we exist to serve. But prohibition is a poor answer to a legitimate grievance, and adopting it would cost us something we are not willing to spend: the artist’s right to decide how they work.
This is the part of the argument that matters most to us, and the part that gets the least airtime.
CUBE’s founding commitments include bilingual programming, fee-free access, and partnership with organizations serving people this town’s institutions have historically reached last. In that context, a tool that translates instantly is not a novelty. Neither is one that lets someone without research training find their way into an archive, or lets a person pressure-test an idea by arguing with something at midnight when there is no one else to ask.
For a great many people, the barrier to making art was never talent. It was access—to training, to studio time, to a language, to somebody who would take the work seriously. These tools lower some of those barriers meaningfully, and they open the door to people who were never going to be let through it otherwise. A prohibition raises those barriers back up, and raises them highest for precisely the people CUBE was founded to reach.
Our commitments, so that this is a policy and not a mood:
On the same terms as any other work: is it good, is it grounded in the artist’s own intent and judgment, does it belong in front of this community.
For CUBE grants, commissions, and exhibitions, artists will be asked what tools they used and what they themselves decided, shaped, and made. A sentence or two—not a confession. We are not policing process. We are giving the audience back the thing that generative AI took away: the ability to know what they are looking at.
Not because AI touched it, but because human authorship is our first principle, and an unaltered prompt output is not a work of authorship. Federal copyright law has arrived at the same conclusion, which is convenient—but it is not why we hold it.
CUBE uses AI in its own work—for translation, drafting, research, and for generating rough visual references when briefing a project. Where AI has materially shaped something we publish, we will say so. An arts organization that quietly uses these tools while asking artists to disclose theirs has forfeited the argument.
The law is moving, the tools are moving, and the field has not settled. We will review this position annually and publish what changes.
We know this position will not be universally welcome, including among people whose work we admire and whose partnership we want. We would rather be clear and disagreed with than vague and agreeable.
Our door is open to anyone who wants to argue about it—and that is not a formality. If we have got something wrong here, we would like to know. This document carries a version number for a reason.
Corey Rawdon
Founding Chairman & CEO, CUBE Foundation for Arts and Culture—Version 1.0, August 2026
