Where is our world going if we can no longer trust a photograph? Yesterday, ultimate proof reality. Today, the result of little pranksters zoning out on Midjourney. The answer with our philosopher friends.
Photo caption: BEST FRIENDS_ By ©Julian_ai_art, generated with Midjourney version 5. Start of the prompt: “Barack Obama and Angela Merkel enjoying their best day at the beach together, because they are now retired and no longer have to deal with politics…”.
I believed it, too… I’ve had lunch in front of a cover of Technikart which I find really nice and which is none other than a fake! Yet it was well written on it, FAKE! But no, nothing picked up. It was for issue n°234, September 2019: a little arty and black and white – you had to guess ! Shit. I’m going home, so determined to be done with AI and everything! I cross the Place de Clichy bridge, I sit on the terrace of the Mayday bar, with the sun beating down all afternoon; I’m reading. But all of a sudden, from the total hubbub of the cars piled up there, just behind, I intercept the conversation of the duo seated in front, it’s about AI, again! Am I followed? I listen. For one, it’s bullshit, which will pass, for the other, on the contrary, these ” images are dangerous, the political consequences are already disastrous! “. wow! The conversation takes a turn that scares me, concepts that rain and everything. The now gray sky makes me rush home, that’s too much! Must end. So here I am contacting a philosopher, Pierre Cassou-Noguès, who defends an original thesis according to which machines have a tendency to expand (and why not?), and Grégory Chatonsky, an artist-researcher – yes, it exists! – from the ENS, which has developed a concept that goes against the general imbroglio: imA or artificial imagination. So, is AI dangerous?
“HOW AM I GOING TO INFLUENCE THE STATS AND HOW AM I GETTING IT BACK? » – GRÉGORY CHATONSKY
Author of the first French-language novel co-written by an AI (Internes, RRoses, 2022), Grégory Chatonsky is convinced of the positive impact that AIs can have, particularly in the art world, when we use them. as tools with which it is possible to communicate. ” The question for artists is: how will I influence the statistics and how will I be influenced in return by these AIs? For thirty years, we have all deposited many traces of our existence on the Web. This data has fed recursive neural networks, the famous AIs, which, thanks to statistics, manage to recognize our world and generate data that resembles it. In the industrialization of the media, we are now entering a new stage: the automation of resemblance is, in my opinion, as important as the appearance of photorealism in the 19th ».
GETTING OUT OF THE TURING MODEL
However, AI confuses everyone, and fascinates as much as it worries. In 1954, in The question of technique, Heidegger said to an audience who had just discovered the microwave oven: There is nothing demonic in the technique, but there is the mystery of its essence “: a mystery nestled in the relationship we have with technology. From the first programmable calculating machines of the 1930s, the first mathematical and computer models of biological neurons in 1943 by Warren McCulloch and Walter Pitts, and John Von Neumann and Alan Turing’s demonstrations of the ability of computer programs to perform what we ask them, the AI is now able to reproduce us in any desired SF setting, such as having Macron demonstrate at an anti-capitalist march.
“MIDJOURNEY IS LIKE A PSYCHOANALYSIS OF HUMANITY. » – G. CHANTONSKY.
The democratization of AI, correlated with progress in the calculation capacities of graphic cards and processors, which took a first leap in the 1990s, and a second since 2010, brings us back to ourselves, to our own capacities. Our speed of execution at school or in business seems to be in competition, as if a new world were playing out in which competitiveness will no longer be examined between humans, but between AI and humans. The “Turing test” developed in 1950 by Alan Turing thus consists of a conversational confrontation between a man and a machine, the first having to guess the nature of the second, the second having to deceive, by mimicry with the human race, the first. It is an experience of this kind that Sonia Devillers almost confronted herself on January 18, 2023, by inviting chatGPT to her radio show on France Inter and asking her questions live: in essence, the objective was to show how this AI was able to maintain a conversation with us, while showing that its answers are obtained without regard to the truth but according to the law of the most probable.
LIKE A HAMMER
However, at least for the time of reflection, let’s stop any comparison between AI and ourselves, firstly because it is a battle lost in advance: AI is a technical device whose objective is in substance to surpass ourselves in a specific area, or in several of them and in a coordinated way. Then, because AIs are dependent on our own knowledge databases: they are mirrors of our needs, as well as our determinisms. The question is rather that of understanding to what extent our thinking is built by technology, by machines, by AI. More than just a tool, like a hammer to hit harder and to shape the world according to our own desires and needs, AI can change the relationship we have to the world by shaping our desires, and even our imagination.
A first basis of study is the modification, by AI, of our desires, central theme of The Benevolence of Machines. How the digital transforms us without our knowledge (Seuil, 2022, 336 p.), by the philosopher Pierre Cassou-Noguès. « Machines do not have life, instinct, intelligence, consciousness – on the other hand, they have a tendency to expand by determining our desires “. His idea ? That AI transforms our happiness by conditioning our desires. ” Some applications, such as the AI Cogito Companion, mediate our relationship to ourselves. She listens to telephone conversations and, by analyzing the intonations of my voice, gives me a mood score. Instead of asking myself how I feel, I use my phone to tell me if I was okay. The AI watches me for my good, it benevoles me, just as such an application counts my steps for my good. It is a transformation of our form of life which induces a new relationship to happiness which is no longer in an immediate relationship of myself to myself. “, he concludes.
For our relationship to the world to be upset, AI must not only modify our relationship to desire, but also our imagination. This is the hypothesis of the ImA., of G. Chantonsky. ” The artist explores, drifts and will be in a learning relationship: he teaches the AI about things in the world and he learns from this paradoxical universe. There is a childhood relationship between AI and us, with which we learn to communicate ».
Contrary to the ordinary use that we make of image generators that we control through a prompt, the artist explores to make his imagination work and not commands to leave to the AI the role of the creative. “ It’s funny to see the images that are broadcast in the media: why are these precisely the ones that are broadcast when there are so many other more interesting ones? Because these are images that express the possible! Trump getting arrested is the Oracle of Delphi. The pope is an ambivalent image: the representative of the poor becomes a symbol of wealthcontinues G. Chatonsky. This is the reversible side of our time. I speak of artificial imagination and the imagination of the artificial, because all the images published in the media are symptoms of what we are: they speak hollowly of us. It’s like a psychoanalysis of humanity. »
The status of the “news image” as tangible, irrefutable and direct proof is upset. Like everyone, the “photo” caricaturing the arrest of Donald Trump on March 20, 2023, I first believed in it, because the former President of the United States had announced it to his community a few days later. early. Eliot Higgins, the British journalist author of this media coup, showed by this what is most unacceptable for us today: that the images that we scroll through all day long deceive us. In response, Midjourney banned arrested of his prompt. But the breach is open. Since the appearance of press photography at the end of the 19th century, the image is our guide to truth; she now seems to be like her shadows projected into the Cave of the Republic of Plato: signs of our alienation to their dictatorship.
By Alexis Lacourte
#Decryption #war #clichés