An AI-generated print sold for $432,500 at Christie’s auction house in New York on Thursday, over 40 times more than it was expected to fetch.
The print is called Edmond de Belamy and its washed-out features—artifacts of the algorithm used to generate it—make it look a bit like the Spanish fresco depicting Jesus Christ that was disfigured so badly in a botched repair job that it became a meme. The print is one of a series of 11, all AI-generated and depicting the members of the fictional Belamy family, and is signed with the mathematical formula describing the algorithm that was used to generate it.
“Behold the future—here it is,” the auctioneer declared before the bidding started on the piece, which is the first AI-generated artwork to be sold by a major auction house.
Far from being the sole creation of an AI, the piece is really the product of months of work by three guys living together in a Parisian flat—one of whom is a machine-learning PhD student—who collectively call themselves Obvious.
The piece’s inclusion in the Christie’s auction, next to prints by Chuck Close and Jeff Koons, has been the cause of some consternation in the art world, but also among AI experts who take umbrage with the implication (by virtue of the signature on the piece) that an algorithm can be an artist all by itself—especially the relatively humdrum variety that was used to create the piece, “Generative Adversarial Networks (GANs).” GANs were developed in 2014.
“The algorithm isn’t the only thing that went into creating this art—GANs don’t have free will,” Mark Riedl, an associate professor of AI and machine learning at Georgia Institute of Technology, told me over the phone. “They’re really, really complicated paintbrushes with lots of mathematical parameters, and you can use this paintbrush to achieve an effect that might be hard to achieve otherwise.”
This nuance was clearly lost in some of the reporting around the piece leading up to the auction—numerous headlines described the piece as being “created” by AI, as opposed to generated, or made with.
Read More: This Algorithm Makes Its Own Video Games
When I spoke to Hugo Caselles-Dupré, a machine learning PhD student and one-third of Obvious, over the phone, he chalked this characterization up to “sensationalism” and “clickbait” in the media. The intent of the piece, he insisted, is to actually educate the public on the limits of AI. Algorithms are a tool, Caselles-Dupré told me, not creative beings themselves.
“Today, it’s not about algorithms that are replacing people,” Caselles-Dupré said. “In the future, we might have to be careful about this, but today, they’re more like a tool. We really wanted to showcase a concrete example of what these AI tools can do.” Signing the piece with the algorithm’s mathematical formula was a “funny way,” he said, to communicate these ideas to a general audience.
It’s unclear how well that messaging strategy has worked. In a statement to Artnet Richard Lloyd, Christie’s international head of prints and multiples said that the piece was selected for auction precisely because of how supposedly little human intervention went into creating it.
GANs are algorithms that “learn” from a large amount of input data, and use that knowledge to produce new results after a long training period. But just because they “learn,” GANs are not autonomous. The end product is the result of a long process of carefully selecting input data, tweaking mathematical parameters, and then sifting through the results to find the very best examples of whatever it is you’re looking for.
“We had to choose the data and curate it manually.Then we had to train several times,” Caselles-Dupré told me. “You have to choose a set of parameters, and then you check the results manually. Like, ‘OK I like this result, so maybe I’ll change the parameters a bit to get something a bit better.’ We did this over a lot of iterations.”
The final iteration of the algorithm—the best it was ever going to get for Obvious’ purposes—spit out hundreds of images, Caselles-Dupré said, which had to be whittled down to just 11. “We carefully selected the images that we found the most interesting in this batch,” he said.
So, which is the really creative party in this process: the algorithm that needs to be prodded and cajoled for months to turn out something half-interesting, or the people searching for an aesthetic result and making all of the decisions to get there? While he work is something of a collaboration between the machine and the people involved, the balance of creativity falls on the side of humans.
“It’s much more about the human and the tool than just the tool itself,” Riedl told me. “My research pushes against those boundaries, and I don’t think we’ll ever get the human fully out of the loop.”
Get six of our favorite Motherboard stories every day by signing up for our newsletter.