Artists feed themselves at all times on technological innovations; sometimes these are determinant and disrupt art. Especially when they come at the right time, like the apparently modest development in 1841, by the American painter John Goff Rand, of the soft metal tube of paint, compactable, hermetically closed with a clamp, which will enable artists to paint outside, leading thirty years later to the first impressionist exhibition.
Today, everybody talks, undiscerningly, about artificial intelligence. Without understanding anything about it, some see in it an absolute intelligence, able to free men from boring tasks and even to create works of art; others see in it a monstrous intelligence, which will distrust men and which, as soon as it can, will get rid of them. Hence the urge, say some of them, to master it before it destroys us. And for other to decrypt what would be an “artistic intelligence”.
We are far, very far away from that. And even if AI already knows how to learn by experience, recognize faces, drive a car and play go better than men, we are very far from seeing it master what constitute the basis of human intelligence. For example, whereas a five-year-old child recognize what is a dog once he saw one, thousands of attempts are needed for a device to learn how to recognize such an animal. And a lot more will be needed for it to possibly be able to create original pieces of art.
Indeed, machines are very far away from being able to compete with men. Let’s check the facts: there is in our cortex about one hundred billion of neurones, all connected, and each of them connected by thousands of synapses; which makes billions and billions of connections. Excluding those which are elsewhere in our body.
The main part of intelligence plays out in the mutual relations between the neurones. It is not in the ability to organize causalities, but to anticipate, by retroactive loops, later perceptions: a tennis player can’t send the ball back after a serve if he only trusts what he sees, a slalom skier doesn’t pass the gates if he doesn’t anticipate the movements of his body. This anticipation is peculiar to the intelligence.
It is hard to imagine that AI could soon compete with that, and it is particularly true for the artistic dimension of intelligence: indeed it is in this anticipation that there is at least a part of aesthetic pleasure, as Francis Wolff showed so well in his book “Pourquoi la musique ?” (Fayard éd.).
Could we say that artists possess a particular intelligence? Will it be possible to approach it by using the rudiments of AI, to write music, to paint, to sculpt, to create? Some people will, already do in self-serving: for instance, to predict the nature of works that the market would want to acquire. Or to protect intellectual property by the interplay of AI and blockchain.
On a more positive note, we can already approach the reality and the complexity of artistic intelligence by the use of virtual reality in order to put ourselves in other persons’ shoes, to see the world differently or to wander in the brain’s labyrinth. Knowing that we’ll only find what we are not looking for, which is the characteristic or true discoveries that matter and real works of art.