Ah, artificial intelligence—the tech everyone loves to hate. Indeed, while A.I. has proven itself loveable as a tool that could aid or enhance creativity, it is also deserving of every side-eye it gets for the real threat it poses to artists’ rights and livelihoods. That’s not to mention the amount of A.I.-generated slop that now clogs news feeds and Google image searches, fueled by Big Tech’s increasing investments in the field.

In 2024, as in every year since generative A.I. entered the chat, the tool’s outsized impact on life, work, and art continued to be unpacked, sparking ever more debate and scrutiny over its potential and peril. Here, we revisit how the A.I. conversation played out this year in museums, on social media, in the courts, and in our headlines.

A.I. Fails

Promotional poster for Civil War (2024), shared on A24’s Instagram account. Photo: @a24 on Instagram.

This year, while A.I. was put to service in pretty impressive projects like a “digital twin” of St. Peter’s Basilica, it elsewhere continued to set off alarms. The producers of the horror film Late Night With the Devil found this out when they included a handful of A.I.-generated frames in their movie, as did production house A24 when it opted to let an algorithm design the posters for its 2024 release Civil War.

And, in what was either a troll or just an entirely unwise artistic decision, an X user deployed A.I. to “fill in” Keith Haring’s Unfinished Painting (1989)—a work the artist intentionally left incomplete to represent the loss wrought by the HIV epidemic. Aptly, the move was deemed “disrespectful” and “disgusting,” with artist Molly Crabapple telling Hyperallergic that such a use of A.I. is “a way for witless dullards to suck every bit of anima, pathos, and humanity out of art.”

Keith Haring, Unfinished Painting (1989). Photo: © Keith Haring Foundation

More stunningly, the London Standard, rather than generating or reanimating art, used A.I. to resurrect a whole person. Or, more specifically, his critical voice: in October, the U.K. paper made waves for publishing a review A.I.-generated to mimic the style of its late critic Brian Sewell (the issue’s front page bears a similarly generated portrait of Prime Minister Keir Starmer). The exercise, as our valiant critic Ben Davis pointed out, was more “cheapo provocation” than anything, failing to confront A.I.’s impact on arts journalism in any meaningful way.

Illustration from Infinitive Wonderland in the style of John Tenniel.

Davis also caught A.I. art as it degenerated into slop over the year, via Samsung’s integration of A.I. photo features into its new Galaxy phone and Google Labs’s Infinite Wonderland project, which served up some nonsensical takes on John Tenniel’s art. They’re damning outputs of Big Tech’s rush to sink dollars into the hot new sector—a phenomenon best indicated by the presence of one Mark Zuckerberg, previously seen renaming his entire company for the metaverse. His Facebook platform, as Davis recently noted, has become ground zero for a creepy new form of “nature slop.”

A.I. Wins

Refik Anadol. Installation view of Living Archive: Nature (2024). Photo courtesy of Refik Anadol Studio

Which is not to say that artists did not find creative ways to wield A.I. Refik Anadol, digital art star and longtime A.I. evangelist, unveiled his first Large Nature Model in January. Built on data from sources including the Smithsonian Institute and National Geographic, as well as techniques from LiDAR to photogrammetry, the open-source model can generate fantastical images of nature, which Anadol showcased in the installation Living Archive: Nature (2024) in Davos and at the Serpentine Galleries show “Echoes of the Earth: Living Archive” in London.

Meanwhile, French collective Obvious debuted its mind-to-image project, presenting an A.I. that can generate images based on a user’s imagination. One of the first fruits of the initiative, an image titled Stagnant Elixir’s Sweet (2024), sold for a tidy $28,000 at Christie’s.

Rendering of Ai Weiwei’s Ai vs AI on the Piccadilly Lights in London. Photo: © CIRCA.

In January, Chinese artist Ai Weiwei also made his first foray into the field with Ai vs AI, a work that blanketed London’s Picadilly Circus with A.I.-generated responses to his philosophical questions. Later, Cai Guo-Qiang, furthering his ongoing inquiries into alien intelligence, unleashed a fireworks display, inaugurating PST Art in Los Angeles, designed in collaboration with his A.I. model cAI™. Before the spectacle, he declared the tech “a revolutionary tool and a tool for revolution.” (Unfortunately, some  of the neighbors were none too pleased.)

Finally, as you’re reading this, performance artist Alicia Framis is officially wed to her A.I. hologram boyfriend Ailex. Congrats to the happy couple.

A.I. on the Block

Ai-Da, A.I. God (2024). Photo: courtesy Sotheby’s.

Obvious joined digital artist Botto, aka Mario Klingemann, on the auction block this year to vend their A.I. wares. All, however, were overshadowed by non-human artist Ai-Da, whose A.I. God (2024) fetched a cool $1.1 million at Sotheby’s. The triptych, a tribute to computer pioneer Alan Turing, is the first work by an A.I. robot to sell at auction; Ai-Da’s mastermind, gallerist Aidan Meller, stated that the robot’s images “question where the power of A.I. will take us.” To the bank, maybe?

A.I. authenticates

Albrecht Dürer, Vna Vilana Windisch (1505). Courtesy of Diego Lopez de Aragon.

This year, Swiss firm Art Recognition made news for authenticating a Dürer drawing, before identifying some 40 forgeries that were being sold on eBay. These determinations speak to the promise of authentication by A.I., even if, as the case of de Brécy Tondo taught us, the approach still has a long way to go to instill confidence. The Swiss authenticator seems ready to take it head on through: “Now, more than ever, it is imperative to stress the significance of adhering to rigorous scientific standards,” said the company’s CEO, Carina Popovici. “Otherwise, the entire field of A.I. could face criticism, and we would all suffer the consequences.”

A.I. vs. artists

Images generated by text-to-image tool Stable Diffusion. Photo: Stable Diffusion.

“It would be impossible to train today’s leading A.I. models without using copyrighted materials”—so claimed OpenAI in a statement to the U.K. House of Lords in January. It’s a remarkable admission, though unsurprising to the many artists who’ve had their work mined and scraped to train these large language models. In fact, in August, a California court validated these fears when it allowed a class action lawsuit brought by a group of visual artists against Stability A.I., Midjourney, DeviantArt, and Runway A.I. to move ahead. Most damningly, in the ongoing case, the judge found that the Stable Diffusion model was “created to facilitate that infringement by design.”

Outside the courts, artists found myriad ways to fight and denounce infringement by A.I. Nightshade, a tool that protects artists’ works by “poisoning” generative models, which was launched earlier this year to resounding success: the software was downloaded more than 250,000 times in the week it went live. An activist group even hacked Disney’s Slack in an apparent protest against the company for “its approach to A.I.” In October, a whopping 11,000 artists, musicians, writers, and actors signed an open letter condemning the training of A.I. on creative works as “a major, unjust threat.”

OpenAI announced it would launch its A.I. video generation tool Sora to the public this year. Photo: CFOTO/Future Publishing via Getty Images.

More recently, Sora, OpenAI’s shiny new video generator, had its source code leaked by artists who were given early access to the tool. About 20 creatives scorched the company in an open letter accompanying the leak, decrying that they were “lured into art-washing.” Their participation, they added, was “less about creative expression and critique, and more about PR and advertisement.”

But perhaps the most understated yet slyest action came from photographer Miles Astray, who snuck a non-A.I. image into an A.I. competition—and won. Astray’s bronze award was rescinded once his feint was revealed, but his point was made. “Winning over both the jury and the public with this picture,” he said in a statement, “was not just a win for me but for many creatives out there.”