The “paintings” produced by AI experiments are not paintings. There’s no paint.
They are simulacra:
“Simulacra are copies that depict things that either had no original, or that no longer have an original.”
Engineers and computer scientists have little real experience or deep understanding of art other than what they see around them and they are using the language of art without understanding it. Artists and educators also don’t know what these things are as they often have little understanding of deep tech as a medium.
We need a new set of words and a new aesthetic to describe this sort of work. My own work is currently around the production of technically generated simulacra, but with the hand of the artist extant in the production.
My own work (instagram.com/richardfadams)is about creating simulacra using computation and digital means to allow me to exploit the language of european expressionism in order to describe my angst and existential terror, as a way of dealing with PTSD, while sitting inside, and tearing apart, selfie culture.
When an AI is trained to make a “painting” it’s following rules setup by a human and generates images according to decisions it makes based on those rules. It surely can expand outwards as it explores, but the images are not paintings they are copies of which there is no “original”. If you put an Infinite number of painters in a room, with the same rule set, they would similarly produce all such variants. But they would be paintings as they would be made of paint. The only conscious aesthetic in these works is man-made, the machine itself does not have an aesthetic. It may have preferences that emerge but it has not consciously made those choices. There are a few variation of the definition of “aesthetic” and these vary depending on the context, however they all refer to an aesthetic being the rules that emergent properties we call art are generated from. Humans set rules in AIs and their work emerged from that. This work is a human piece of work produced for them by a smart machine. The works produced are by definition simulacra.
If you think this distinction is not really important then ask yourself what the machine would make if it had its own aesthetic.
AIs in this context are simply speeding up and exhausting the artistic process of try, fail, try again, set question, test question and try again. This is yet another clever manifestation of the new aesthetic of push-button art.
Ps. If anyone is interested in hearing me talk more about this in the context of how educators need to change art education, then i’ll be talking about it at the Valencia Design Education Forum in November.
#art #design #creativity #AI #machinelearning