March 2022

Arts & Letters

Market of the apes: NFTs and digital art

By Richard Cooke

Bored unique limited edition NFT monkey art by Mininyx Doodle / Shutterstock

NFTs have transformed the art market, but artificial intelligence might transform art itself

“For the delight of his spirit and the joy of his eyes, he had desired a few suggestive creations that cast him into an unknown world, revealing to him the contours of new conjectures, agitating the nervous system by the violent deliriums, complicated nightmares, nonchalant or atrocious chimeræ they induced.”

— Joris-Karl Huysmans, Against Nature

In April 1965, the Howard Wise Gallery in New York hosted one of its least successful shows. Called Computer-Generated Pictures, its title was a compromise between the two men whose work was on show. They were scientists rather than artists, and they disagreed about whether the images they had created could be called “art”. Béla Julesz thought they were not art, and A. Michael Noll believed they might be, but they did agree on one inevitability: “the day when a computer can draw – or paint – almost any kind of picture in any one or combination of colors”. Invitations to Computer-Generated Pictures were printed on coloured IBM punch cards; one went to a New York Times critic who declared that “the wave of the future crashes significantly at Howard Wise Gallery”. But it was low tide: of the 25 works displayed, some of which might have been made on an Etch A Sketch, none sold.

The written material that accompanied Computer-Generated Pictures offered reassurances, which sounded empty in the face of market indifference. Artists who feared being automated out of existence might instead find a freedom “unburdened by the tedium of the mechanics”. Feeling more inhibited, executives from the telecom company AT&T tried to cancel the exhibition, fearful that regulators might punish an expensive waste of computing power. Noll was undeterred, and his experiments became more provocative. He reverse-engineered a Mondrian work, and created an algorithmic “original” Mondrian with a computer. When he showed the results to test subjects, they preferred them to the real thing. It was more “soothing”, respondents said, and more “imaginative”. Noll, whose programming had made “no attempt to communicate any emotions”, was fascinated.

He then tried to register an image called Gaussian-Quadratic with the Library of Congress copyright office, which rejected it. “A machine generated it,” the library responded, beginning a bureaucratic dialogue about whether or not a computer could be original. “I explained that a human being had written the program that incorporated randomness and order,” Noll wrote in his short memoir, The Beginnings of Computer Art in the United States. “They again refused to register the work, stating that randomness was not acceptable. I finally explained that although the numbers generated by the program appeared ‘random’ to humans, the algorithm generating them was perfectly mathematical and not random at all. The copyright was finally accepted.” Gaussian-Quadratic became perhaps the first known registered piece of copyrighted art produced with a digital computer.

The work lasted decades as a novelty – almost a gimmick – rather than a harbinger. Thirty years after its creation, its creator had still “not noticed any major museum of modern art giving much attention to computer art”, mostly, he thought, because it was not very good. Instead, it has taken almost 60 years for the possibilities Noll and Julesz explored to reach fruition. A lot of computer art is still not very good; nevertheless their central question – will artists be emancipated or replaced by digital technology? – has become a much more urgent consideration. And like so many of the art world’s crises, this one is finding expression at the auction house.

Christie’s has handled art made with computers and art made by computers for some time, but in March 2021, it sold its first ever piece that exists on computers and nowhere else. This purely digital artwork is called Everydays: The First 5000 Days, and the artist responsible calls himself Beeple. It was minted exclusively for Christie’s as a non-fungible token (NFT), and comprises a collage of 5000 drawings that Beeple posted online, one per day, for 13 and a half years. Everydays was part of Christie’s debut sale of NFTs, the first of its kind by a major auction house. It also marked the first occasion that the 256-year-old institution accepted cryptocurrency as payment. The work sold for the equivalent of $69 million, ranking it among the most expensive artworks by living artists. High prices for modern art have been tabloid fodder for nearly a century now, but “money for nothing” was an irresistible new variation, and the sale launched a flurry of both awed and despondent headlines.

Prices like these provoke anxiety, especially when attached to objects that have no corporeal existence. NFTs are unique units of data, their provenance proved by lines of code on a blockchain. An artwork produced as an NFT exists only in the digital realm. The work is copyable, but the ownership isn’t. Some have suggested the copyrighting and sale of digital assets is not really new at all: video art, for example, has been sold in limited editions for decades. Brian Eno dismissed NFTs as artists’ own “cute little version of financialisation” that allowed them to “become little capitalist assholes as well”. Trevor Jones, a 51-year-old painter turned digital artist from Edinburgh, had more reason for gratitude. Before NFTs, he told The Guardian, he had been borrowing money to pay bills; afterwards, he made more than $4 million in a single day. The journalist noted that though this transition had transformed Jones’s bank account, his status still rested near anonymity.

Much of Jones’s work is adorned with crypto-currency logos and paraphernalia. If these are intended ironically or as a critique, his fans pay it no mind. Contemporary galleries found these artworks objectionable and, instead, buyers arrived through online marketplaces such as Nifty Gateway and OpenSea, where speculation is feverish. Combined with a pandemic-induced decline in museum and gallery attendances, this amounted to a reversal of fortunes between traditional and non-traditional art-world institutions. In an ArtReview piece titled “Why the Artworld Loves to Hate NFT Art”, the critic J.J. Charlesworth argued that “pop-cultural baubles” had become a viral threat to the established order. What did they mean for institutions? What did they mean for art criticism? “After all,” Charlesworth wrote, “to this critic, Beeple’s and Grimes’s images suck.” Beeple himself had described much of his own work as crap. It didn’t seem to matter – the eventual owners of Everydays admitted they had not previewed the lot before purchase. If it was crap, it was crap they had never seen.

Some critics reached for their Marx, who wrote in Capital that “all nations characterised by the capitalist mode of production are periodically seized by fits of giddiness in which they try to accomplish the money-making without the mediation of production process”. (In the same discussion, the philosopher noted that the accumulation “appears in a form that leaps to the eye”.) But there is a sharper, less distant manifestation of this phenomenon: the art market of the 1980s, a time when, as Robert Hughes described it in The Shock of the New, an “overstressed decade” became a “bullring of deranged fetishism”. Stratospheric prices mocked the idea of art as a socially shared medium and threatened to destroy it altogether. The art world subordinated itself to television, with its “ultrafast change of images, its throwaway cool, its predilection for banal narrative, and its fixation on celebrity”.

Artists such as Jeff Koons made work that was nasty but not cheap. It flaunted its own ephemerality and ran open lines of communication with fashion and street style. There was closer proximity to celebrity, and a diminished role for critics – Hughes’s own rearguard against a trend that he hated was unsuccessful. Then, as today, the whole hot, gauche period was buoyed by financial sector exuberance.

If NFTs have a Peggy Guggenheim, it might be Paris Hilton, who owns more than 150 digital works, some displayed on screens in her house. She has produced a series of her own, depicting both her dog and herself in a fairy floss, airbrushed style. In August 2021, she discussed NFTs on The Tonight Show Starring Jimmy Fallon (she was invited back for more), and some months later Fallon purchased an NFT of a cartoon ape, minted by the art collective Bored Ape Yacht Club (BAYC), for 46.8 ethereum, a cryptocurrency, then worth around $250,000.

Not the first or the most expensive NFTs, BAYC apes may be the most emblematic. Distinct from self-contained NFT works such as Everydays, the BAYC is a collection of collectibles, exactly 10,000 cartoon-style portraits, programmatically generated at random from a selection of different traits. Accoutrement such as glasses, or golden fur, are distributed throughout the primates; each has a predetermined rarity, which increases the value of the ape accordingly. All 10,000 apes sold one day after the project launched, in what The New Yorker called “a strange combination of gated online community, stock-shareholding group, and art-appreciation society”. (It could have added streetwear brand, baseball-card company and money laundering service.) “We want your Bored Ape to be your digital identity,” Gargamel, one of BAYC’s anonymous founders, told the interviewer via video chat. Those who own these ape images, Fallon included, frequently use them as social media avatars.

BAYC was inspired by an NFT project called CryptoPunks, and has in turn spawned its own menagerie of imitators and sub-projects. There are Desperate Apewives, the Fat Ape Club, the Mutant Ape Yacht Club, along with a host of other copycats producing cartoonish images of pugs, penguins, lions, ducks, whales (and cats). This ballooning animal kingdom has been compared to the Beanie Baby Bubble of the 1990s, where fortunes sank on ultimately unsaleable plush toys, but there are outright scams as well as delusions. In late 2021, both the Baller Ape and Evolved Ape NFT projects (the latter run by a developer known as “Evil Ape”) suffered multimillion-dollar “rug pulls”, where their originators disappeared with the money. There have also been numerous thefts: New York gallerist Todd Kramer was the victim of one of the first digital art heists in history, when phishing thieves stole $15 million worth of BAYC NFTs. His defeated tweet in the aftermath – “all my apes gone” – has become one of the slogans of a disorientating zeitgeist.

Applying art criticism to algorithmically created cartoon apes, stolen or otherwise, might be a kind of category error. Superficially, they sit in a lineage with Keith Haring and KAWS, at the juncture between art, street art, graffiti, collectibles and graphic design. But then context runs aground. Those artworks were produced by individual human artists, and even the most cynical progenitors of such works still had something to say, even if it boiled down to “buy me”. Here, it is the medium that makes that demand. As with Noll’s programming, there is “no attempt to communicate any emotions”, and any feeling a work might produce is an accident, a kind of emotional pareidolia. There is no connoisseurship, and no pretence that those buying the apes are part of a liberal humanist effort to preserve the human soul.

Noah Davis, the Christie’s specialist who has become an impresario and interpreter of digital art markets, believes that NFTs exist in opposition, designed “against the existing art world” as an “an art form that doesn’t need a gallery”. That suggests that the “art” lies both within and beyond its electronic superstructure. When more than 1 trillion copyrighted images are being created each year, individual works of art have a better chance at being rare than transcendent. The end point is a place where the market itself is the gallery, or even a new form of networked, collaborative artwork. It’s a phenomenon that seems related to other online communities that revel in deliberate financial “irrationality”. The Reddit forum wallstreetbets, where users post “loss porn” of heavy trading deficits and urge each other to commit to doomed stocks like GameStop, is also an aesthetic collaboration. At their best, these created spaces continue the emphasis on waste and mis-purposing found in so much performance art.

Works bought and sold without the involvement of any “real life” artists deepen the longstanding unease surrounding digital artworks and their artifice. Negative reviews of Computer-Generated Pictures called the show cold and soulless; both human and algorithmically generated digital art still attract these pejoratives. “What distinguishes Beeple’s digital imagery from other ‘non-establishment’ art is the violent erasure of human values inherent in the pictures,” Jason Farago wrote in The New York Times, “and how happy his cryptofans are to see them go.” That doesn’t seem so distinguishing – Futurism was a whole gospel of anti-humanism, and Paul Virilio was so incensed by Stelarc’s man-becomes-machine installations he likened them to Nazism – but the sheer volume of digital works exist on a scale that is counter-Romantic. Whether image number 4999 of a series is moving or well executed becomes immaterial: the viewer never made it that far.

Perhaps the most groundbreaking modes of digital art embrace this proliferation instead of resisting it. Amid the postmodern retreads and reinventions of 1980s SoHo, Robert Hughes found a tired déjà vu “born of the total availability of all images”. This abundance was only relative (Hughes had seldom seen colour reproductions of Old Masters before leaving Australia), but the comprehensive computerised archives of our own era offer themselves as a medium instead of an inspiration. These images are machine readable, and subject to the deep learning at the core of contemporary artificial intelligence. Computers trained on these enormous banks of imagery can function in ways that recall thinking, or dreaming, and just as they can recognise faces, or make rudimentary medical diagnoses, they can also make art.

In another of A. Michael Noll’s less well-known pieces of writing, “The Digital Computer as a Creative Medium”, written in 1967, the scientist imagined computers acting as an artist’s assistant or creative partner. Both a tool and a medium, they could be “used to produce wholly new art forms and possibly new aesthetic experiences”. If originality came from the unpredictable, a machine “actively takes over some of the artist’s creative search. It suggests to him syntheses that he may or may not accept. It possesses at least some of the external attributes of creativity.” This was a prescient guess at how AI technologies would come to be used: Sougwen Chung, an artist-in-residence at Bell Labs (Noll’s former employer), has described her work as “mark-making in collaboration with a robot”.

This artistic movement is sometimes called network art or neural art (one artist-in-residence at the Google Cultural Institute in Paris called himself a “neurographer”), because of the nature of the artificial intelligence preferred. Many practitioners use computers that function like a neural network, modelled on a brain. In particular, artists favour a form of artificial intelligence called generative adversarial networks. GANs use two intelligences working in tandem: one generates images and one classifies them, a relationship that might be likened to an artist and a critic. In 2015, Google released a project called DeepDream, which aimed to visualise what a deep neural network was “seeing” as it functioned. It has since become a favourite of AI artists, and the imagery it creates – much of it in an overdetermined, psychedelic style – has been loosely collected into an artistic movement called Inceptionism.

Inceptionist works are vivid and fractal, and the DeepDream AI tends to see animal faces everywhere, even on top of other animal faces. Tonally, they can be nightmarish and sinister – the first work produced by generative adversarial networks was a serendipitous creation the internet named Nightmare Beast. Asked to visualise dumbbells, DeepDream attached disembodied arms to each one. The human artists who work with these tools have tamed some of these uncanny effects. The Turkish new media artist Refik Anadol, who happened to be a Google artist-in-residence present for the release of DeepDream, has since applied similar techniques in public art works all over the world. At the end of 2021, he held a show at New York’s MoMa called Unsupervised: “an exhibition of works created by training an artificial intelligence model with the public metadata of The Museum of Modern Art’s collection”.

Aside from its presence in hallowed institutions, neural network art also has a lively grassroots scene. It is cheap to produce and easy to learn, which makes it democratic, and frees a creative impulse from the strictures of craft. The musician and producer Vito James Genovese reinvented himself as a DeepDream artist, producing a steady stream of work shown in as many as three online exhibitions a month. Each show is dedicated to art-historical styles of his own invention. They have suggestive names – Ultra Deco, Rainbow Doom, Century Style – and at their best suggest a route to new artistic territory, much like Giorgio de Chirico’s arcade paintings.

Ultra Deco imagines what artists of the 1920s might make with digital tools – “the availability of digital color saturation would be a major marvel for those early 20th century art deco guys i think” – and Genovese takes a limitless pleasure in working the controls of his AI. He finds source imagery from old film clips or found footage, and finesses the textual prompts that tell the machine how to think. They can be technical and descriptive – “teal purple silver color scheme:50 | palm tree vacation beach paradise:50”, for example – but sometimes the prompts become poetics. “The fidelity of AI is what interests me so much,” he says. “How one powerful prompt can bring with it the feeling like some singular artist is replicating the same general feeling, yet it’s all just the work of words and algorithms.”

The analogy Genovese likes to apply is of synthesised music, specifically the experimental form that began in earnest in the 1960s before commercial synthesisers broke it into the mainstream. It often sounded cheesy or artificial before it changed the sound of music permanently. The closest parallel between the machine’s role in both music and art is the ability to generate nearly anything synthetically. Up close, Genovese says, the machine’s personality feels unmechanical. “The inferences and ideas that get drawn out from input images especially are quite mesmerising. The way the AI ‘understands’ is almost intimidating.”

The artist Casey Reas, author of Making Pictures with Generative Adversarial Networks, describes this as coaxing images from “latent space”, and affirms that it’s still an artistic process. He compares curating the images from which an AI learns with a photographer’s practice: choosing the best marriage between equipment, subject and conditions. At each step the artist must make decisions, and the mystery of the computer’s function makes this process organic and still humane. Still, an AI artist is not a photographer, and Reas is honest about the nagging questions of intent and purpose, and how both critics and audiences might respond to them.

Mike Pepi, who writes about art and technology, has tried to imagine the critic’s role in the AI art world. “A critic writing about DeepDream could, for example, point to the narrowness [in its] selection of images of cats and dogs, and how the end product suffered as a result.” The critic might parse and compare output, but that role would “treat the artist-engineer as a curator, as someone whose primary work is the selection of images”. This is because the true moment “of machine creativity is illegible to humans. If it weren’t, then there would be nothing artificial about its intelligence.” In her own work, the curator Nora N. Khan has imagined AI art as an ethical practice that could undermine the surveillance regime of digital capitalism, addressing and subverting its algorithmic regimes.

Once, when Genovese fed his AI a random waterscape image, he twinned it with the prompt “seaside structure by nighttime ocean”. He was imagining a bridge; instead, the computer created giant abstract sculptures out near the sea, and added a rural shoreline in the foreground. Looking closer, Genovese saw something that looked like strips of tilled farmland and, on the horizon, the twinkling of a city skyline, elements that existed only in the mind’s eye of a machine.

Richard Cooke

Richard Cooke is The Monthlys contributing editor. 


From the front page

Composite image showing John Hughes (image via Giramondo Publishing) and the cover of his novel The Dogs (Upswell Publishing)

A dog’s breakfast

Notes on John Hughes’s plagiarism scandal

Image of Erin Doherty as Becky Green in Chloe. Image supplied

App trap: ‘Chloe’

‘Sex Education’ writer Alice Seabright’s new psychological thriller probing social media leads this month’s streaming highlights

Pablo Picasso, Figures by the sea (Figures au bord de la mer), January 12, 1931, oil on canvas, 130.0 × 195.0 cm, Musée national Picasso-Paris. © Succession Picasso/Copyright Agency, 2022. Photo: © RMN - Grand Palais - Mathieu Rabeau

‘The Picasso Century’ at the NGV

The NGV’s exhibition offers a fascinating history of the avant-garde across the Spanish artist’s lifetime

Cover image of Paul Dalla Rosa’s ‘An Exciting and Vivid Inner Life’

‘An Exciting and Vivid Inner Life’

Alienations and fantasies of escape unify the stories in Australian author Paul Dalla Rosa’s debut collection

In This Issue

Photograph of Jacqui Lambie

Goddamn bloody adult: Jacqui Lambie

The Senator Lambie reality show

An photograph of the spacewarping effect, taken with the NASA/ESA Hubble Space Telescope, and depicting GAL-CLUS-022058s, located in the constellation of Fornax.

Founding NEMO

Gravitational waves and the secrets of the universe

Illustration by Jeff Fisher

Car sick

When the satisfaction of being your own mechanic turns to the unease of driving a car you have put together yourself

Photograph of Gabrielle Chanel at her home in Faubourg Saint-Honor, Paris, 1929.

Beyond the little black dress: ‘Gabrielle Chanel – Fashion Manifesto’

The NGV showcases the controversial designer’s undeniable brilliance

More in Arts & Letters

Robert Fielding, Western Arrernte and Yankunytjatjara peoples, ‘Graveyards In Between’, 2017

Standing and ceremony: The 4th National Indigenous Art Triennial

Themed around ‘Ceremony’, the NGA exhibition provides a moving examination of what it means to be Indigenous in 2022

Image of James Joyce and publisher Sylvia Beach in Paris

The consecration: James Joyce’s ‘Ulysses’

A century after its publication, the difficult reputation of Joyce’s seminal novel has overshadowed its pleasures

Image of Tom Cruise, circa 1980

Sixty business: Tom Cruise

‘Top Gun: Maverick’ hits screens as its ruthlessly career-oriented star turns 60

Image of Steve Toltz

The quip and the dead: Steve Toltz’s ‘Here Goes Nothing’

A bleakly satirical look at death and the afterlife from the wisecracking author of ‘A Fraction of the Whole’

More in Art

Robert Fielding, Western Arrernte and Yankunytjatjara peoples, ‘Graveyards In Between’, 2017

Standing and ceremony: The 4th National Indigenous Art Triennial

Themed around ‘Ceremony’, the NGA exhibition provides a moving examination of what it means to be Indigenous in 2022

Mirdidingkingathi Juwarnda Sally Gabori, ‘Dibirdibi country’ (2008

Art heist: The landmark conviction of an Aboriginal art centre’s manager

The jailing of Mornington Island Art’s chief executive for dishonest dealing has shone a light on ethics and colonialism in the Indigenous art world

Detail of Anne Wallace’s ‘She is’ (2001)

A new kind of history: ‘Know My Name’

A radical exhibition of women artists at the NGA draws wondrous connections across time and between disciplines

Image of René Block installing 7000 Oaks on behalf of Joseph Beuys as part of the 1984 Biennale of Sydney

From little things: 23rd Biennale of Sydney

The story behind the 40-year-old tree centring ‘rīvus’, José Roca’s Biennale

Online exclusives

Composite image showing John Hughes (image via Giramondo Publishing) and the cover of his novel The Dogs (Upswell Publishing)

A dog’s breakfast

Notes on John Hughes’s plagiarism scandal

Image of Erin Doherty as Becky Green in Chloe. Image supplied

App trap: ‘Chloe’

‘Sex Education’ writer Alice Seabright’s new psychological thriller probing social media leads this month’s streaming highlights

Pablo Picasso, Figures by the sea (Figures au bord de la mer), January 12, 1931, oil on canvas, 130.0 × 195.0 cm, Musée national Picasso-Paris. © Succession Picasso/Copyright Agency, 2022. Photo: © RMN - Grand Palais - Mathieu Rabeau

‘The Picasso Century’ at the NGV

The NGV’s exhibition offers a fascinating history of the avant-garde across the Spanish artist’s lifetime

Composite image of Sydney Morning Herald editor Bevan Shields (image SMH/supplied) and actor Rebel Wilson (AP Photo/Chris Pizzello)

Two sides of the same Shields?

Editor Bevan Shields’ attempts to handle the backlash over his masthead’s treatment of Rebel Wilson points to the dismal and fragile state of news media