Artificial Intelligence starts making art. But, is it art? What is art?
AtParatissima Art Fair in Turin last week, the Feral Horses Team decided to launch a quiz to understand people’ interest and knowledge about Artificial Intelligence (AI) and get the chance to win a share from JAGO. To do so, we put together a list of fourteen artworks created by both artists and computers and then asked people to guess who made what. Interestingly enough, pools results had been extremely constructive and it made us realise that it is not always easy to guess the real author of an artwork. Following the quiz, we came to the very question for the participants… Can we consider art those artworks made using AI? Once again, the opinions were very mixed, and doubt was king! But at the very end, does it really matters? Shouldn’t it be the pleasure derived from an artwork, rather than knowing if it is machine-made or not, that brings people to buy a piece of art?
AI rising fast!
It is on everyone’s lips that AI art is gaining popularity in the last few years. Some gallery across the globe started exhibiting AI art and, to mention only one, a major exhibition in New Delhi Nature Morte Gallery was organised earlier this year, being ‘the first ever art exhibition in India to include artwork made entirely by artificial intelligence’. To those very into AI and art, some of these names of the artists and coders participating to the exhibition might ring a bell, namely Mario Klingemann, Memo Akten, Tom White and Anna Ridler.
Not everyone seems happy about the sale
But as we know success never comes alone, and in this case it came with a lot of noise! Artists such as those listed above, expressed strongly their frustration and complaints about Portrait of Edmond Belamy, first work of AI art auctioned at a big auction house.
The Christie’s lot was produced by the French collective under the name Obvious — which works using generative adversarial networks (GANs). For those who are interested in the technicality of the GAN, here it is how it works: Two algorithms compete to perform training. And then “a generator creates new images by mimicking characteristics of images from the training dataset, and try to fool a discriminator into thinking those images are “real”. The generator trains until no difference can be made by the discriminator.
Put simply, these systems involve two programs, one of which holds a database of something real (in this instance, pictures of old paintings) while the other produces images that try to “fool” the first into thinking that they constitute an example of what it knows. Gradually, the second program develops ever-more convincing (but completely invented) images in the “style” of the database examples”.
“Obvious stole the spotlight with work that is seen as basic and derivative” was only one of the concerns showed by many of the pioneers in AI art.
In their defence, the small but passionate group of AI artists explained that “there was no paint or brushes involved” and that it was only an algorithm that did all the work — in this case, thousands of portraits spanning the 14th century to the 20th”.
Not everything that is written is truth
The reality might be a bit more complicated than what it seems. In fact, as explained by Hugo Caselles-Dupré, technical lead at Obvious, in an interview with ArtNome, people have completely misunderstood the full story behind the Portrait of Edmond Belamy. He affirms that Obvious was only trying to make their first project original, without pretending at any moment that machines could make their own art without a human behind, and the intention to be under the spotlight of Christie’s sale.
There is no doubt that we need to give some credit to both sides and for now the only things that need to be kept in mind when thinking of this painting (and of Obvious) are:
· “A human chose the data set
· A human designed the network
· A human trained the network
· A human curated the resulting outputs”
· AI can create super cool stuff!