The Artsy AI Survey 2026: What Galleries Really Think About AI In The Art World Artsy

(MENAFN- USA Art News) Artsy's 2026 AI Survey Finds Galleries Embrace AI for Operations, Not as an Artistic Medium

Artificial intelligence has become a quiet fixture in the commercial art world, but not in the way many technologists might expect. According to Artsy's inaugural 2026 Artsy AI Survey - a poll of more than 300 gallery professionals published March 18, 2026 - galleries are already leaning on AI for routine business needs while remaining notably skeptical about AI-generated work as a category of contemporary art.

The results sketch a market in transition: AI is increasingly normalized behind the scenes, yet debates about authorship, originality, and labor continue to shape how dealers, artists, and collectors respond to the technology.

A definition problem at the center of the debate

One of the survey's clearest findings is also one of its most basic: the industry still lacks a shared vocabulary. More than a quarter of respondents (28%) said their gallery does not have a formal definition of“AI art.”

Among those who do define it, the answers diverge sharply. Some respondents (22%) described AI art as fully prompt-based or generative work in which the primary composition is produced by AI. Others (18%) offered a medium-agnostic definition that prioritizes artistic intent over the tool used. Another 16% classified any work in which AI“meaningfully shapes the outcome” as AI art.

That lack of consensus matters because it shapes everything downstream - from how galleries talk to collectors to how they position works in a program, and even how they evaluate value. Without a stable definition, the market is left to negotiate legitimacy in real time.

Skepticism about AI as a medium remains widespread

While artists working with AI - including Turkish American artist Refik Anadol (b. 1985), German artist Mario Klingemann (b. 1970), and Chinese Canadian artist Sougwen Chung (b. 1985) - have achieved institutional visibility and market traction, the survey suggests that many galleries still hesitate to treat AI-generated art as a fully established medium.

Only 9% of gallery professionals surveyed said they consider AI-generated art a legitimate new medium. A larger share framed it in more uncertain terms: 25% described AI art as a“destabilizing force” for authorship and value, and 28% called it an“evolving category” whose market value remains unclear.

Those concerns echo broader anxieties about how AI systems are trained. Because many models draw on vast datasets that can include copyrighted artworks, questions of consent and compensation continue to shadow the technology's adoption - particularly in a field where provenance, attribution, and the economics of labor are central.

Collector demand is muted, with pockets of curiosity

If galleries are cautious, collectors appear even less engaged. The survey found that 41% of galleries said AI“rarely comes up” in conversations with collectors. Another 16% reported that collectors“actively avoid” works made with AI assistance.

There are signs of interest, but it is often exploratory rather than transactional: 15% of galleries said they have encountered“curiosity-driven interest,” in which collectors ask questions without necessarily buying.

The findings suggest that, for now, AI art is not a consistent driver of sales within traditional gallery channels - even as attention to AI-generated imagery grows elsewhere online.

Most represented artists are not using AI in their studio practice

The survey also points to limited adoption among artists represented by the galleries polled. Sixty-one percent of respondents said none of their artists currently use AI in their practice. Nineteen percent reported that one to two artists they work with use AI tools, while 8% said three or more of their artists incorporate AI.

Where AI is used, it tends to appear in specific, practical applications rather than as an all-encompassing method. The most common artistic uses reported include rendering and visualization (48%), AI-enhanced photography or image-making (47%), training models with personal datasets (44%), and research or conceptual development (39%).

Artists are also using AI for professional support tasks: writing or editing bios, statements, and CVs (36%); image editing and file management (27%); administrative writing (25%); and studio organization or project planning (21%).

A market learning to live with AI's contradictions

Taken together, the survey's portrait is less about a sudden AI takeover than a slow recalibration. AI is already embedded in the infrastructure of galleries and studios, but the commercial art world remains wary of what it means for authorship and value - and unconvinced that AI-generated work has secured a stable place in the market.

As the technology advances and institutional attention continues, the next phase may hinge less on novelty than on clarity: clearer definitions, clearer ethical standards around training data, and clearer signals from collectors about what they are willing to live with - and pay for.

MENAFN18032026005694012507ID1110879342