If we take another step backward: what is an artist? And ultimately, what is art?
Let’s begin with this last question:
What is Art?
As I am not an expert in art theory, I will base my opinion on how I understand other experts views on the matter. I know that one cannot condense multiple millennia of art history in a short article, but let’s try analysing AI art in context.
One of the most straightforward definitions of Art that I have heard recently, is the one given by Simon Sinek in his talks. In his very eloquent style, Simon defines the artistic process as the intersection of 3 intentions:
- The intention to create something artistic: write, paint, play;
- The intention to display the creation as art: in a book, a gallery, a concert hall;
- The intention to admire or critique the creation as art;
So in order for an object to be admired as art, it needs these three ingredients, and for each phase, the respective actor (creator, curator, admirer) needs to have the faculty of intention.
This implies that a painting, if it is kept hidden, cannot be considered an artistic object regardless of the mastery of the creation. The same goes for all arts, as creations need to be displayed, heard, admired and critiqued in order for their value not to diminish.
This is what drives large art collectionneurs to display their most valuable pieces in public galleries. Well, except for Madonna and her bathroom Picasso.
Who is the artist in the case of “AI Art”?
The definition above let’s us dissect art and artistic process by identifying the 3 actors. I think that anyone will agree on the fact that the curators and admirers of AI art today are all humans. There is no art gallery, today, curated by AI, as well as there is no poetry book or blog compiled solely by an AI algorithm. So displaying art and admiring art are all human activities.
Join our weekly newsletter to receive:
- Latest articles & interviews
- AI events: updates, free passes and discount codes
- Opportunities to join AI Time Journal initiatives
You might remember from my previous article that we still live in a “weak AI” world. So all the AI techniques available today amount to a set of optimisation tools that augment human potential, but with no intrinsic autonomy. So these tools still need human operators to use or misuse them.
And this is the same with art creation. The intention to create comes from the human pushing the “on” button of a 3D printer on which the next painting or sculpture will take shape. The same for verses and slams, these are all based on human language and poetry rules created by human brains for human ears and used to convey a message.
There are implications to over optimizing a message meant to be received by a human brain. Everyone heard about the Cambridge Analytica story. In a future article, we will look at how art affects the human body and mind and why this is important when it comes to AI art, from deep fakes to mood inducing art.
But let’s not get derailed from our subject at hand.
Will we ever have actual AI art?
The easy answer is: if we ever get to a point where there will be autonomous AI creators, AI curators and AI critiques, humans will not be able to see or understand the result. Remember The Matrix movie, where the matrix simulation was all green source code spiraling down the screen and only some humans were able to “read” it without converters? It will be the same for actual AI art meant for AI brains: there will be no need for a machine to create something other than a succession of bits (1s and 0s) or qubits and humans will be left out.
But this will have other, more down to earth, implications. Let’s imagine an autonomous AI artist able to create streams of “beautiful” bits and qubits that will eventually become its next masterpiece. As every creator, it will need to try its hand and create sketches before beginning the actual work. It will be similar to a child discovering a printer linked to the home computer. Admit it! We all did that at first! With a paint-like software on the screen and the print button at a click’s distance, we were able to finish the paper stack and ink/toner in no time. It will be much the same way for our futuristic AI creator, except that in its case it will all be about hard drive space. I really don’t like the perspective of an entire cloud farm to be filled overnight with AI art sketches blocking out all other operational activities like my favorite online store or AI Time Journal articles. This will be the apocalypse by storage space shortage!
Still not convinced?
Let’s go further and try a complex answer. By combining the two arguments above and knowing that AI techniques are all about optimizing human cognition and intelligence, we may state that AI artistic creation will never be possible. Why? Because the AI artist will never be happy with its work. It will iterate to infinity, trying to perfect its craft and its creation without even getting to the second phase – showing it to a curator. Where human artists are limited by time, money and energy, an AI creator with a constant electricity source will go on creating variants forever, constantly learning and optimizing.
Will leave you with this idea and look forward to reading your feedback.
Product Manager @ Dassault Systèmes