Not surprisingly, mavericks get a hostile reception when they present their innovations to other art world members. Because it violates some of the art world's conventions in a blatant way, the work suggests to others that they will have trouble cooperating with its maker; its blatant disregard of established practice suggests that the person who made it either doesn't know what is right or doesn't care to do what is right.1 Howard Becker, Art Worlds
Christie's auction house is holding its "first-ever," "groundbreaking auction"2 of AI art and it has caused a major stir in the contemporary art world. Many artists who have laboured for years to gain recognition are rightfully concerned about how images of their work have been collected without permission and then used as data to build Augmented or Artificial Intelligence (AI) models. Still others are questioning the legitimacy of this "art" that is produced in whole, or in part, by a machine. Is it really art, and if so, is it any good? These are questions that prodded me into looking more closely at how AI art is valued. Christie’s auction, with its artists, images, and texts, have provided the perfect data set for considering these questions and whether AI art is actually "museum-worthy" as Christie’s likes to claim.
As I have written here before, anything, even a urinal, a chair, a brick, or a can of soup, can become contemporary art if the institutions of the art world recognize the objects or images as art.3 This means that AI art can, potentially, be accepted as “museum-worthy” if these institutions, combined with the people connected with them, authenticate the objects through exhibitions, sales, and the discourses that introduce and interpret them as art. The narrative or story created by these discourses is especially important as it provides the justifications for why the object is "special" or worthy of being art.
Conveniently, Christie's website provides just such a story. The article, "What is AI Art?" is presented as an informative text that is also persuasive. The text is designed to assure the reader that AI art is indeed human. The anonymous writer begins by providing a simple definition of AI art that emphasizes its human origins. AI-assisted art is "any form of art that has been created or enhanced with AI tools." The writer then introduces numerous artists and describes their work, explaining how each artist has used computer data and learning to develop and/or construct the work. According to Nicole Sales Giles, the Director of Digital Art at Christie's (who is also likely the writer of the text), AI art is only "about employing technology to push what is possible, exploring what is achievable outside of, but not separate from, human agency." AI, the article then tells us, offers a "high-calibre" tool for humans that has "boundless possibilities." In the end, the writer assures the reader that "AI is not a substitute for human creativity," but rather a new tool that "enhances the human spectrum of creativity."
While informative, the article is also persuasive in that it provides the justifications for why AI art should be recognized as "museum-worthy." The writer emphasizes how AI art has a "variety of styles" which are a "testimony to the great breadth in character and approach of the artists involved." In order to demonstrate that these many styles are not just a passing trend, the writer builds a history for AI art. A history gives this new art form a solid foundation, one that can be equated to the long history of Western art that contemporary art is founded upon. The writer cites several of the "pioneers" of AI art, such as Harold Cohen (1928-2016) and Charles Csuri (1922-2022), who experimented with early computer programming which then led to the current spate of AI production. AI art, then, the writer proposes "can be traced back several decades" and is actively "creating art history in real time."
But, AI art, like any new art form, cannot convincingly join art history unless it has gained recognition in important art museums in cultural centres like New York, Paris, and London. While some of the AI artists have been producing work for at least a decade or more, their work has been featured in peripheral venues such as galleries (on-line and bricks and mortar) that cater to digital art. These venues do not have the power to consecrate new objects as "museum-worthy." For this reason, the writer is careful to note that some AI artists included in the auction have had exhibitions in major museums. Refik Anadol, for example, has had a solo project at MoMA in New York and Charles Csuri’s work was included in a group exhibition at the Tate in London. And, Holly Herdon and Mat Dryhurst’s work was "the only AI work in the 2024 Whitney Biennale." These citations serve as evidence that AI art can be “museum-worthy.”
Throughout the text, the most frequent value used to justify the worthiness of AI art is the assertion that this new art is "original" and "innovative." For example, the writer notes how Refik Anadol employs a “distinctive and unique approach" while Pindar Van Arman’s use of AI produces "unique canvases." Singular originality is an important value for validating any contemporary art. The writer affirms the singular originality of the AI art presented in the auction by stating that,
"We’re excited by artists who have done deep research, whose work is grounded, and part of a consistent practice, not a one-off. We’re always looking for something we’ve never seen before.”
Like the artistic avant-garde and contemporary art, AI art is also recognized for its revolutionary character: it is "radically contemporary in many ways;" it is "leading to some of the most hyper-contemporary art of the time;" and, it is "art that’s as intriguing as it is disruptive." The only disruption evident in this story of AI art, however, is the artists’ use of AI rather than an aesthetic innovation or a provocative meaning. Van Arman’s “unique canvases,” for example, are “autonomously created by neural networks combined with robotics, a significant milestone in AI-generated art.” AI, the new tool wielded by this “new guard” of artists is what gives their art its revolutionary and provocative edge. As the writer notes,
AI art is so fascinating, because you see artists bending the technology to their will. They’re putting restraints in place, building parameters and then letting AI run free within those boundaries.
The story of AI art that emerges from this narrative is that it is, and should be, valued for its innovative technical presentation that uses augmented intelligence. The writer avoids any discussion of aesthetic quality or the meaning of the work. Thus, unlike contemporary art which is valued for how it addresses contemporary global issues such as identity, climate change, and so on, "AI art is about a new collaborative artistic process with technology."
Justifying AI art for its use of augumented intelligence may not yet be convincing enough to gain it recognition in major contemporary art institutions. But, as Howard Becker points out in his book, Art Worlds, there are always mavericks on the horizon - those artists who "propose innovations the art world refuses to accept within the limits of what is ordinarily produced."4 Often these artists and their work remain on the periphery and are never given any notice in the museums of the world. Others, however, find champions who support their efforts and are then belatedly recognized as innovators who really should have been acknowledged in the first place.
In addition to my analysis of the Christie’s narrative, I also investigated the 34 individual artists whose work is being sold at the auction (you can view the auction site here). The research included searching for information on the artists’ age, education, exhibitions, and documenting their Artfacts ranking. The following chart reveals what I found:
Notes:
· The majority of AI artists were born after 1970, which indicates that most grew up with computers as part of their everyday lives.
· In addition to the human artists, three machines are listed as artists and two of these, like Ai-Da Robot, have Artfacts rankings.
· In terms of education, only six out of 34 artists have completed a BFA or MFA degree unlike most recognized contemporary artists. Instead, most of the artists have studied, or have some training in, computer engineering and coding. Unfortunately, this data is incomplete as many AI artists work under pseudonyms and little information is published on their personal credentials.
· Only one artist, Refik Anadol, is ranked in the top 1,000 on Artfacts.net. This makes him one of the most famous contemporary artists in the world, comparable to artists like Anne Imhof, Peter Doig, and James Turrell, just to name a few. These artists, like Anadol, have had numerous important commissions and exhibitions world-wide, many of them in significant art institutions.
· Six artists are ranked in the top 10,000. All of these artists have taken part in exhibitions in major institutions including the Serpentine, ICA, and Tate Modern in London, MoMA and the Whitney in New York, and Centre Pompidou and Jeu de paume in Paris. This ranking is typical for mid-career artists who are recognized both nationally and internationally.
· Twelve artists are included in the top 100,000 ranking. This is a common ranking for emerging contemporary artists who have gained some respectable exhibitions in their own countries, and participated in at least one group exhibition in international venues or less significant museums. The first four artists in this category, for example, have participated in group exhibitions at the Ullens Centre, Beijing, Louisiana Museum in Copenhagen, V&A Museum in London, and Kunstmuseum, Bern.
· The six artists ranked at 1,000,000 have had at least one solo exhibition in a private or public gallery that has been vetted by Artfacts. This is the same rank as many contemporary artists who have just begun their careers and have never gained international recognition.
· Finally, nine artists do not appear on Artfacts listings at all. This indicates that they have never exhibited in the venues that Artfacts tracks or that Artfacts has not yet received and vetted the information. This does not mean that they have not exhibited at all. Some of these artists have had their work featured on on-line galleries and in small venues that specialize in digital art and NFTs. As the Christie’s story relates, AI art has been around for several decades but circulated in a very different socio-economic field than contemporary art.
· A comparison of Christie’s estimated auction values and the Artfacts ranking indicate that the higher ranked artists on Artfacts are not necessarily the ones that command the highest auction prices. Presumably Christie’s estimated value is based on previous sale figures which might explain why an artist ranked at 1,000,000 might command a price of $20,000-30,000 for their work. The disconnect between the Artfacts ranking and the estimated value of an artist’s work demonstrates how the value of AI art is not, at this time, strongly influenced by an artist’s reputational value (something Artfacts is designed to track) within the contemporary art market. Christie’s hold auctions of many types of collectibles, some of which become just as valuable as important works of art by famous artists. Over time, Artfacts’s data provides a good measure of whether an AI artist’s work remains in the category of a valuable collectible or is transformed and valued as contemporary art.
Howard S. Becker, Art Worlds: 25th Anniversary Edition (Berkeley: University of California Press, [1982] 2008), 233- 234.
All quotes from Christie’s literature are from their website article, "What is AI Art?" and their press release.
See, for example, my post “When is a Chair Art?.”
Becker, Art Worlds, 233.