Only 9% of gallery professionals consider AI-generated art a legitimate new medium. That single data point, buried in Artsy's inaugural 2026 AI Survey, tells you everything about where the commercial art world stands on artificial intelligence: it's a spreadsheet enhancement, not a creative force.
The survey, which polled more than 300 gallery professionals, found that AI adoption is widespread for administrative tasks — inventory tracking, client communications, scheduling — but when it comes to hanging AI-generated works on gallery walls, the industry has drawn a hard line. Twenty-eight percent of respondents don't even have a formal definition of what AI art is. Among those that do, definitions vary so wildly that the category itself remains functionally meaningless.
This isn't skepticism about a new medium finding its footing. This is an industry that has decided AI art doesn't deserve the benefit of the doubt it routinely extends to other experimental practices. The survey's findings reveal a commercial art world that sees AI as operationally useful but culturally illegitimate — a tool for efficiency, not a tool for creation.
The definitional chaos is where the resistance starts. Twenty-two percent of galleries define AI art as "fully prompt-based" or generative works where the "primary composition" is AI-generated. Eighteen percent use a "medium-agnostic" definition focused on artistic intent. Sixteen percent classify any work where AI "meaningfully shapes the outcome" as AI art. The lack of consensus isn't an oversight — it's a feature. Without a shared vocabulary, the industry can avoid having a real conversation about whether AI-generated work belongs in the same ecosystem as painting, sculpture, or photography.
Compare this to how the art world absorbed photography in the 19th century or video art in the 1960s. Both faced institutional resistance, but the resistance was about legitimacy within a shared framework, not about whether the category itself should exist. AI art, by contrast, is being held in a definitional purgatory that allows galleries to use the technology behind closed doors while keeping it off the walls.
The survey's findings on market perception are even more revealing. Twenty-five percent of respondents see AI art as a "destabilizing force" for authorship and value. Twenty-eight percent describe it as an "evolving category" with unclear market value. These aren't neutral observations — they're exit ramps. Calling something "destabilizing" or "unclear" is how the art market signals that it's not ready to bet money on a trend, and in an industry where market value and cultural legitimacy are inseparable, that's a death sentence.
The business logic is straightforward. Galleries operate on consignment models, commission structures, and the long-term cultivation of artist careers. AI-generated art disrupts all three. If a collector can prompt-generate a similar work at home, why pay gallery prices? If the "artist" is a collaboration between a human prompter and a machine learning model trained on thousands of other artists' work, who gets the commission? And if AI tools democratize image creation to the point where anyone can produce gallery-quality visuals, what happens to the scarcity that underpins the entire market?
These are legitimate questions, but the survey suggests galleries aren't interested in answering them — they're interested in avoiding them. The adoption gap between operational AI and artistic AI is the tell. Galleries are happy to let algorithms optimize their back-end processes because that doesn't threaten the core business model. But putting AI art on the walls would require rethinking authorship, pricing, and the entire value chain that connects artists to collectors. The industry has chosen efficiency over evolution.
The irony is that artists like Refik Anadol, Mario Klingemann, and Sougwen Chung have already gained institutional recognition and market traction with AI-assisted work. Museums are hosting AI art exhibitions. Auction houses are selling AI-generated pieces. But the commercial gallery sector — the ecosystem that traditionally bridges emerging practices and mainstream acceptance — is sitting this one out. That's not caution. That's a bet that AI art won't matter in five years, or that if it does, someone else will have figured out how to monetize it.
The survey also highlights a deeper tension about what galleries are actually selling. The art market has always traded on the myth of the singular artistic vision — the idea that a work's value derives from the irreplaceable hand and mind of the creator. AI art, by design, complicates that myth. If a neural network trained on millions of images produces a work, who is the creator? The person who wrote the prompt? The engineers who built the model? The thousands of artists whose work trained the algorithm? Questions of authorship and attribution are already causing legal headaches in other industries, and galleries would rather not import that chaos into a market that depends on clear provenance and ownership.
But the refusal to engage with AI as a legitimate medium also reveals something about the gallery world's broader cultural conservatism. This is an industry that prides itself on being at the vanguard of contemporary culture, yet when faced with a technology that could fundamentally reshape how images are made, it defaults to gatekeeping. The 9% who consider AI art legitimate are outliers, not early adopters. The 28% who see it as destabilizing aren't wrong, but their response is to reject the disruption rather than adapt to it.
The survey's findings also connect to a larger pattern in how cultural industries are responding to AI. Hollywood is experimenting with AI for production workflows while keeping it out of creative decision-making. Music labels are using AI for A&R research but not for songwriting. Publishing houses are deploying AI for editing and marketing but not for manuscript generation. Across the board, the strategy is the same: adopt AI where it enhances existing processes, avoid it where it threatens existing power structures.
For galleries, the power structure at risk is the artist-dealer-collector triangle that has defined the commercial art world for centuries. AI art doesn't fit neatly into that triangle because it challenges the scarcity model that makes the whole system work. If image generation becomes as easy as typing a sentence, the market for unique, hand-made objects doesn't just shrink — it becomes a fundamentally different category. Some galleries will adapt. Most, according to this survey, would rather wait it out.
The question the survey doesn't answer is what happens when the 9% who are engaging with AI art start outperforming the 91% who aren't. If AI-assisted artists develop new aesthetics that collectors want, if museums continue to validate the medium, if a younger generation of buyers grows up seeing AI as just another tool in the creative toolkit — then the galleries that rejected it early will be playing catch-up in a market they helped create. That's the real risk the survey reveals: not that AI art will destabilize the market, but that the market will destabilize itself by refusing to evolve.
