Aydın and Agoria, the art and way of creating with AI

Can Dall-E and Midjourney turn everyone into an artist? To answer, it is enough to look at the way and approach of Aydın collective and Agoria.

Spectators flocked to the Danysz Gallery this evening on December 3rd, lined up in front and at the entrance, despite the chilling and majestic but unsettling sculptures of American Mark Jenkins for some, risking scaring passers-by. by. Founded in the heart of the Marais by art dealer Magda Danysz at the end of the 20th century, the space hosts Obvious “7.1”, a new exhibition by the French collective. It is, among other things, about artificial intelligence in the service of art, a concept that is still unclear to many people.

While presenting his work to the assembly, a colleague tells them that he is trying to create images through artificial intelligence; in fact, he can also claim to be an artist. Hardly embarrassed by this response, stating that one brush in hand is enough to paint a masterpiece, the collective patiently responds. A priesthood. “Current photographers don’t have to explain how their cameras work; soon this will happen with artificial intelligence. A good part of our speech is educational and explanatory, especially about the place that is always important in the creative process. by assimilation we will be able to get to the heart of the subject we are dealing with,” – Gautier Vernier, one of the three members, admitted to us a little later.

Since the creation of the collective, the three protagonists of “Aydyn” have been dealing with misunderstanding and even mistrust. Soon to be documented in a documentary on Canal+, their rise to fame came almost by accident in 2018 when one of their paintings with features from the intelligence program was auctioned at Christie’s in New York. Edmond de Bellamy, which sold for more than $430,000, made headlines and caused a stir. “We were completely overwhelmed by the scale of events,” recalls Pierre Fautrel, one of the co-founders. To the boiling point of leaving New York and rallying Paris on the same night.

AI will reimagine African masks and Japanese prints

Always on point, Obvious has never stopped producing, always relying on the same approach: the last member of the trio feeds Hugo Caselles-Dupre’s algorithms with organic, human inspiration. The “7.1” exhibition, presented at the Danysz Gallery until January 14, consists of canvases depicting the seven wonders of the world, which have now disappeared, except for the remains of the Pyramid of Giza. To realize them, the collective, with the help of a historian, fed the algorithm with reference to a collection of ancient texts grouped together and then presented visual reinterpretations of these images. Works put on canvas by a studio of artists selected by the collective also remain virtual through augmented twins that are distributed on NFT and visible after scanning in Artivive software.

In 2020, with “AGI Facets”, Obvious offered a series of African wooden masks with features derived from the code and designed by a Ghanaian artist. A few months ago, the collective asked AI to design eleven scenes that were repeated in prints from the Japanese Edo period. With the inhabitants of Ile-de-France, History always provides the basis of the method. “Our message wants to end the duality of output, whether it’s predictions of machines replacing humans or the growing enthusiasm for technology. Art and science are connected, it’s a constant evolution, and we show that for these Artists, a new tool, a new creation Our style and collections present occasional analyzes of new technologies. It is never black or white,” continues Pierre Fautrel.

A few days after this interview, Sébastien Devaud, a good acquaintance of the collective, also known by the pseudonym Agoria, confirms these ideas. After a very successful but secret meeting for 2023, he meets us in a cafe on the left bank, smiling. The Nuits Sonores de Lyon co-founder has just come off a long evening at Transmusicales in Rennes, during which he offered the 2,500 people in the hangar a chance to strike a snapshot of the visual creativity at NFT and the music produced during the evening. A first. “I wanted to attract an audience that isn’t necessarily Web3-friendly, which is an alternative,” the producer tells us. “This is Jean-Louis Brossard (Founder of Transmusicales de Rennes, editor’s note), very interesting, very open, who ordered this creation from us. It was good to offer them, we do not want a trade relationship in this context. It’s a permanent memory on the blockchain.”

Ambitious as it may be, this experiment in NFT was far from the first for Agoria: together with coder friend Johan Lescure, he has been experimenting with NFT support and algorithms for several years, especially “since the pandemic”. Like last year’s Obvious, it’s infused with organic matter and digested by machines. Phytocène, composed in 2021 with Nicolas Desprat and Nicolas Becker, reproduced in music and images the data collected by probes of the symbiosis between a plant and its bacteria. In April, he collaborated with biologist Alice Meunier for a series of NFTs that show on video the moment the human brain makes a decision.

Only Sébastien Devaud refuses to use the term artificial intelligence. “Nobody wants to deal with anything artificial except maybe a few paradises,” he justifies. “I prefer to talk about augmented intelligence or algorithmic intelligence.” Like his obvious friends, he considers this technology “a boon to the imagination.” He says that every morning he thinks about the fruit of the expression of neural networks. “There are endless possibilities as a result of algorithmic queries. And you can rewrite it every day. It’s dizzying.”

“What questions does a tool like Dall-E lack of work”

Agoria smiles again when he tells the anecdote of the “Aydyn” team, which was questioned by our colleague mentioned above. “He’s right,” she replies without irony. “I think AI exists as a tool, and they question the place of the artist first and foremost, always putting him on a pedestal. With AI, it’s important that the artist creates his own databases: if you start with existing data, as Dall-E does, the approach becomes simpler, with a careful It has to be. People won’t be fooled. It’s not about what can be easily produced, it’s about whether it makes sense. It has to be work. At the end of the day, a tool like Dall-E questions is not work.”

However, more than five years after Obvious’s inception, is the art still relevant in an era of algorithmic popularization? Back at the Danysz Gallery, Pierre Fautrel, Hugo Caselles-Dupre and Gauthier Vernier say nothing but Agoria, who prepares thousands of tests for each work before agreeing on the final result. “This evolution is not something we are afraid of, we were supporters of this creative revolution and it is happening today,” answers the first. “Actually, we’re glad we didn’t miss the mark. Our art form has always been conceptual, driven by a message, and it’s great that it took five years for technology to democratize. Now for good ideas and creativity. Anyone can buy a brush, canvas, and paint, but anyone can be an artist. No. We still have something to say.”

Next year, the trio will begin opening a research lab, a dream that will come true thanks to groundbreaking work by Hugo Caselles-Dupre, a doctoral student in machine learning at the Sorbonne. “Today we use the technology, but now we want to participate in its development,” explains Gauthier Vernier. For its part, Agoria is already working on many projects, for example, the follow-up to the latest NFT produced in collaboration with Ledger and the Renaissance agency, or other achievements that combine algorithm and bioscience.

So, if the coming year already promises to be fruitful, how do they see the medium-term future, as indicated by their respective presence at NFT Paris in February? They don’t risk prophecies if they mention text-to-video “for next year, by the end of five years” and brainwave interpretation. “This will have a significant impact on many professions, especially creative ones, but projecting yourself is very complicated. But it’s a mental exercise that we often do,” Gauthier Vernel admits. “We have a 100% chance of cheating on this answer!”