Artificial intelligence is arriving in art and professionals in the sector are wondering how to manage this new data which is moving the lines of artistic creation.
The Dall-E 2, Midjourney and Stable Diffusion programs wowed the planet last year with their ability to invent images from a few elements of text, and works produced with artificial intelligence were sold for thousands of euros. Today, artistic circles are still looking for the right position in the face of the unprecedented image-creating capabilities of new artificial intelligences (AI), believes Emily L. Spratt, art historian, specialist in Byzantine and Renaissance art but also expert in AI.
The lure of novelty
Typing for example “Brad Pitt in a canoe in space, in the style of Mondrian”, brings up in a few seconds a very colorful image of the American star paddling in the stars. So, Dall-E 2 and consorts defend the idea that they make it possible to democratize art by giving everyone the power to create. But it’s a vision “far too simplistic (And) naive” believes the art historian. These tools are above all “a way to push the use of large internet platforms, which of course is very good for these companies”. “In the future, artificial intelligence will complete the entire architecture for making digital images”, she continues. It will mix with other technologies for manipulating images, already widely used, analyzes Emily L. Spratt.
Especially since digital works produced with artificial intelligence already have a market. Some have sold for tens and sometimes hundreds of thousands of dollars. Among the inventive artists, the expert cites the German Mario Klingemann. His work Series of hyperdimensional attractions, bestiary is a video of “seemingly organic shapes that constantly evolve to briefly resemble this or that well-recognizable animal”, she explains. Mario Klingemann’s art raises the question of artificial intelligence as a means of expression, and more broadly that of the sources of creativity, summarizes Emily L. Spratt.
Selective sorting
Three 29-year-old childhood friends, gathered in the collective “Obvious” (“obvious” in French), pioneer of the genre, have already bet on the flight of this artistic revolution. In 2018, the French collective had made an impression by selling in New York, at Christie’s, a work for more than 400,000 euros (“Edmond de Belamy“), explaining that she had been “created” by artificial intelligence. This sale bristled some artists specializing in AI, because the Obvious collective admitted that it was not the author of the algorithm which made it possible to create the image. The latter, Robbie Barrat, was not indignant at this use of his code.
“If Obvious’s work sold at this price, it was largely because it had been presented as the first work produced by an artificial intelligence, and sold by a major auction house”explains Spratt. “It was really the art market that was trying the experiment to see what a work produced by artificial intelligence could offer, while following traditional sales methods”, she explains. Since then, the reversals of fortune in the tech world have dampened enthusiasm. Both Christie’s and Sotheby’s have created separate platforms to sell these works, “as if”says the historian, “they didn’t want to sully art with these new digital explorations”.