It is a question of articulating creation and human history. Hugo Caselles-Dupré, a member of the Obvious collective, explains the role that artificial intelligence could have in the world of creation by making new works.
Hugo, could you introduce the team and your background?
The team is made up of three childhood friends. After each following a different path, especially in economics, entrepreneurship and artificial intelligence, we decided to continue our fascination with technology and especially for artificial intelligence by embarking on an artistic project.
What was your goal in creating Obvious?
The desire to embark on this project struck us when we discovered the existence of THE Generative Adversarial Networks algorithms capable of generating entirely new visuals by analyzing a very large number of examples. This algorithm fascinated us, and we thought that these technologies deserve to be known to all. Our goal was to make the understanding and use of these technologies accessible.
Why did you choose art as a field of intervention in relation to artificial intelligence?
In our opinion, art is the best way to address the world’s population universally. It is not limited by language, and everyone has a personal and intimate relationship to this universe. We thought this was the best way to get our message and vision across. Moreover, art is a particularly interesting field of experimentation when working with artificial intelligence, a playground with infinite inspiration and which is not governed by any rules. The idea of combining science and art seems particularly interesting to us because it represents the greatest creative challenge for a machine.
How do you differentiate yourself from other digital art creations?
We immediately decided to take a different approach than most artists who work with technology for the creation of art. We wanted to create works that resonate in the collective imagination as classical works, so that the viewer could relate to something known, and thus has a point of comparison. In our view, this was the best way to show the state of AI research and the discoveries that had been made.
Can you describe the process of creating a work?
The creation of a work begins with the choice of a subject, which must above all correspond to the message we wish to convey. We then learn about this subject by seeking to meet the various experts, and by conducting research by ourselves. We then collect visual data, through partnering with institutions, or by recovering images free of law when the subject allows us. After sorting this data, we encode the algorithm or use open-source versions of it to behave in relation to that data. This process is often tested and error-based, between which we make changes to the code and the database. Once we arrive at a satisfactory result, we select the visuals that best serve our approach and our message. We then look at the material production of the works, trying to best match the codes of the artistic movement we treat.
How did the idea come for your first collection of 11 portraits of the Belamy family?
We thought about what we wanted to convey: algorithms capable of being inventive were discovered. We have selected art as the best way to express it. So we thought of the most universal art form, and the most obvious to us. Probably skewed by growing up in France and Europe, we naturally turned to classical portraits.
How many portraits or data did it take to get to Edmond de Belamy’s portrait?
We used 15,000 portraits, which range from the 14th to the 20th century.
There is a lot of debate about the design of a work of art calculated by an algorithm. How do you react to criticism?
We are very grateful to have been critical because they mean that we have managed to start the debate that we wanted to share. We respond mostly by destroying prejudices about how artificial intelligence works, to get critics to see AI as yet another artistic tool, certainly fascinating in its intrinsic capacity to create, but also as a tool that does not yet have any intention or self-awareness.
Since the spectacular sale at Christie’s, what has happened to your projects?
We have released a new series of 22 Japanese works, bringing a new generation of GANs on a very large number of Japanese prints, representing portraits or landscapes. We have also participated in a large number of events to present our approach, and we have worked with brands such as Nike, HBO, Serviceplan, and we are approached by institutions such as the Quai Branly Museum, Semitour which manages the heritage of the caves of Lascaux, or the Museums of Paris as part of a project to reinterpret their respective collections. , and the creation of new pieces made by the AI. Finally, we have presented our works in various exhibitions in prestigious museums including in Quebec at the Museum of Civilization, in the United States in Florida at the Art Fort Lauderdale, in China at the National Museum in Beijing, in Russia at the Hermitage Museum in St. Petersburg, and in Saudi Arabia at the King Fahd Cultural Centre in Riyadh.
©Obvious, Bulma of the High Red Pines
What are your future plans?
We are currently working on our next collection called Facets of AGI, which will consist of 22 African masks created using the GANs. This series talks about Generalized Artificial Intelligence, and the avenues of research to reach it, drawing parallels with a fictional secret society that venerates intelligence through various ritual masks. We are also working on a pair of Nikes whose colors have been determined by an algorithm, via the Nike By You Workshop project. We are also working on various long-term projects involving 3D, creating a new Marianne using artificial intelligence, or allowing artificial intelligence to dream.
Follow the work of the Obvisous collective on their website.
Interview by Eleftheria Kasoura