Close-up of signage for Christie’s auction house in Manhattan, New York City, New York, September 15, 2017. (Photo by Smith Collection/Gado/Getty Images)
Getty Images
Christie’s, a 258-year-old institution synonymous with fine art, is about to step into what it calls a “new frontier”: AI-generated art. The “Augmented Intelligence” auction is scheduled to open for bidding online and at Christie’s New York on Feb. 20 and runs through March 5. But instead of applause, it has been met with outrage. Thousands of artists have signed a petition calling for the cancellation of its AI-art auction. The controversy is not just about AI, it is about ownership, authorship, and what art even means in the digital age.
The petition, which quickly gathered over 6,000 signatures, argues that many of the AI-generated works being sold were created using models trained on copyrighted works without permission. It further criticizes the legitimization of AI-Generated Art in the fine art market, warning that this auction sets a dangerous precedent where technology companies profit from creative labor without compensating the original artists.
Who Owns AI Art?
The core argument against Christie’s AI auction is that many generative AI models have been trained on vast datasets that include copyrighted work, often without consent. This is not a new issue, artists have long protested against generative AI companies scraping their works without credit or compensation. But now, one of the world’s most prestigious auction houses is seemingly endorsing AI-generated works by selling them at high-profile auctions. The question is: Can an artist ethically sell a piece of AI-generated work if the AI itself was trained on a vast, unauthorized collection of other artists’ creations?
According to a Christie’s spokesperson, the auction house claims that the works in this sale are “in most cases” trained on the artists’ “own inputs.” But that phrase raises more questions than answers. AI models cannot be effectively trained on just one artist’s work alone. These models require massive datasets, millions, if not billions, of images, to function. Given the complexity of AI training models, there are questions about how much transparency buyers have regarding the origins of these AI-generated pieces. Additionally, for collectors, the debate extends beyond provenance to the evolving perception of value in the art market.
A diptych from Holly Herndon and Mat Dryhurst’s “xhairymutantx.” The work, originally commissioned as part of a larger piece for the 2024 Whitney Biennial, is among those offered for sale at Christie’s landmark AI art auction.
Christie’s Images Ltd. 2025
The Copyright Minefield
The US Copyright Office has repeatedly ruled that pure AI-generated works are not eligible for copyright protection. If that remains the case, what exactly are buyers at Christie’s purchasing? Without copyright, anyone could replicate the same AI-generated work, meaning there is no “original” to own in the traditional sense. In the fine art world, uniqueness is everything, so does this sale fundamentally devalue the entire concept of collecting?
Some critics argue that Christie’s is not just legitimizing AI-generated art but actively devaluing the rest of its collection. If AI-generated pieces sell for tens or hundreds of thousands of dollars, how does that impact the value of traditional, human-created art?
What is Christie’s Endgame?
Christie’s has sold AI art before, notably, in 2018, when it auctioned “Portrait of Edmond de Belamy” an AI-generated painting that fetched $432,500. But this is its first auction dedicated entirely to AI art. Why now? Is this about embracing a “new medium,” or is Christie’s simply testing the waters to see how much profit AI-generated work can bring in?
The timing is important. The creative industries are in a heated battle over the impact of AI on jobs, copyrights, and ethics. Hollywood writers recently fought against AI encroachment in their contracts. Musicians are challenging AI-generated deepfakes of their voices. Visual artists have been on the front lines, battling AI models that have trained on their work without consent. And in the middle of this cultural and legal war, Christie’s decides to host an AI art auction.
The AI Art Divide
Christie’s has defended its decision to auction AI-generated works. “AI is learning everything that it possibly can from an entire set of data and images to then create something new,” said Nicole Sales Giles to NPR, Christie’s director of digital art. “That’s influence. Not theft.” However, many experts argue that this misrepresents how machine learning works, AI does not merely ‘learn’ from datasets; it deconstructs, analyzes, and recombines elements from vast amounts of human-created art, often without consent or monetization. This process goes beyond mere influence; it appropriates the creative output of countless artists, raising serious ethical and legal concerns. Critics contend that presenting this as “influence” rather than “unauthorized appropriation” is a fundamental misrepresentation of how generative AI operates, fueling ongoing disputes over intellectual property rights.
Giles also stated that copyright issues surrounding commercial AI models trained on artworks featured in the sale are beyond the auction house’s purview. This position distances Christie’s from the growing legal challenges around AI-generated art, including cases like The New York Times suing OpenAI for copyright infringement.
Giles also highlighted the increasing demand for AI-generated art. According to a report from Business Research Company, which forecasts the global AI art market to reach nearly $1 billion by 2028, doubling its estimated value from 2023.
Not all artists are against the auction. Some, like Refik Anadol and Mat Dryhurst, have embraced AI as part of their creative process. They argue that AI is merely another tool, much like photography was in the 19th century. Dryhurst, for instance, dismissed the backlash as “doomsday hysteria” in The Guardian, pointing out that his AI-generated pieces explore the concept of publicly available images.
But critics counter that there is a difference between using AI as a tool and selling works generated by ethically questionable models. The AI models in question were not built in isolation; they were trained on an internet’s worth of human-made art. If an artist “curates” AI outputs but does not program the AI itself, is that still art? And if the AI-generated work is indistinguishable from human-created pieces, does that signal a creative breakthrough or a creative collapse?
The Ethics of AI Training
The bigger issue at play is how AI is trained. If artists had control over whether their work could be used to train AI models, this controversy might not exist. But today, most artists have no such control. Some AI developers argue that training on public images falls under “fair use,” while others are actively working on ethical AI models that allow artists to opt in. Until then, AI art remains a Wild West of intellectual property concerns.
Maybe a possible solution: If AI art is to be sold, the artists whose work was used to train the AI should receive royalties. This would at least acknowledge the contribution of those who, unknowingly, helped shape these AI models. But without enforceable legal frameworks or transparency, such an approach remains theoretical.
Buyer Beware?
Perhaps the most ironic twist in this debate is the possibility that AI-generated works auctioned at Christie’s might be valueless in the long run. With AI progressing at lightning speed, today’s groundbreaking AI-generated pieces may soon look like outdated digital relics. And given that AI-generated works cannot be copyrighted, what stops someone from using the same model to create nearly identical pieces and sell them elsewhere?
The Future of AI Art in the Fine Art Market
Christie’s decision to hold this auction may be a watershed moment for AI art, but whether it is remembered as a pioneering step or a colossal misstep remains to be seen. What is certain is that AI in the creative industries is not going away. The industry will have to grapple with questions of authorship, ethics, and value, whether it likes it or not.
One thing is clear: The backlash against Christie’s AI auction is not just about one sale. It is about who gets to profit from art in the age of artificial intelligence, and who gets left behind. And until those questions are answered, every AI-generated artwork sold at auction will be a piece of a much larger, unresolved debate.