Is the Generation of Images with Artificial Intelligence Art?

Ernst Gombrich already said it in his immortal work History of art: There is no such thing as Art with a capital letter, but rather artists. The eminent historian based his assertion on the fact that each era and each culture has had its own concept of what art is; Not only that, but each individual also has their own ideas about it, so we can conclude that no, art in the abstract does not exist.

However, there is a common denominator that links the art made by all communities in history: and it is precisely that, the fact of being linked, exclusively, to human beings. Until very recently it was unthinkable to conceive art as something not linked to human nature; You could discuss what its characteristics were, what was essential to be considered “art”, but it always emerged from a human hand.

Currently, this idea has undergone a change, not without controversy, by the way. And in recent years we have witnessed the rise of so-called artificial intelligence and everything that this entails: from greatly facilitating the development of our daily lives to generating works that aim to emulate the art that has been created for millennia. The inevitable question is: Is creating images with AI really art? This is the topic we will delve into throughout the article.

Can creating images with AI be called art?

If art is linked to human expression, we can impulsively affirm that no, this type of “creation” cannot be called art. But, first, let’s define a little what artificial intelligence (AI) consists of and the “art” that is developed through it.

Artificial intelligence and the GAN

Artificial intelligence or AI (AI in its English acronym) is a set of technological systems and machines that aim to emulate the processes followed by the human brain, with the aim of facilitating work. At least, this would be its definition in its beginnings, because currently the possibilities of action of artificial intelligence have expanded to unsuspected limits.

Specifically, AI has multiplied its branches into the field of art. The invention, in June 2014, of the Generative adversarial network technology (GAN for its acronym in English: Generative Adversarial Network). Those responsible for the invention were computer scientist Ian Goodfellow and his team, who devised an algorithmic system organized around two antagonistic neural networks: a generator and a discriminator. These two networks compensate each other: while the generator creates images from a very extensive database, the discriminator “validates” the creations through an identification process.

This revolutionary technique allowed the French artist collective Obvious to generate the first “work of art” made with artificial intelligence This is the famous painting Edmond de Belamy which, by the way, owes its name to the surname of one of the creators of the GAN technique, since bel ami In French it can be translated as “good friend”, more or less the same as Goodfellow.

The work was auctioned at Christie’s in October 2018 and sold for the not inconsiderable figure of $432,500, even more impressive considering that initial valuations priced it for only $10,000.

A new artistic reality?

He Edmond de Belamy of Obvious It is not, by any means, a crowning work made with this technology. Since its appearance until now, many more images have been generated with the GAN technique that have turned out to be much more convincing. In any case, the true importance of Belamy is that, for the first time in history, placed a work created with artificial intelligence in a prestigious auction house and also multiplied its starting price This makes us wonder: is artistic reality changing? Is the future of art inevitably linked to this type of creations?

It is undeniable that artificial intelligence has greatly enriched films, advertising and video games, but this technology entails dangers of which it is necessary to be aware. It is very likely that you have heard of the deepfakes (literally, deep, grandiose falsehoods), in which well-known people appear performing actions that they have never actually performed. In this type of fakesfor example, a perfectly recognizable actor or singer appears designed through artificial intelligence.

Nowadays, these types of techniques are still in full development and it is relatively easy to discover the deception, but What will happen when technology improves to such an extent that it is impossible to distinguish truth from lies?

The “mechanical” art

But, ethical debates aside, let’s return to the basis of this article, which is to consider whether or not images created with artificial intelligence are art. It is clear that the fact that this type of creations are slowly becoming established does not justify their appreciation as artistic manifestations. We have already commented that, for a work to be considered “art”, it must be created by a human mind. It is true that behind artificial intelligence there are human beings. But, ultimately, it is the machine that generates this type of creations, through algorithms, so it is risky to call these types of objects “works of art.”

In general, it is considered that the artist must “bring out” what lies within him; that is, his work is an expression of his soul. However, this statement also requires some nuances, since Not in all times or in all cultures art was something inherent to the “soul” In the Romanesque, for example, painters and sculptors were mere artisans, who limited themselves to reproducing what they were asked for and what they were paid for. A Romanesque artist hardly put all his expression in a work, since each of the motifs that he had to capture were duly codified.

In other words, he had clear instructions on what to paint or sculpt and how to do it. This does not mean, of course, that a certain freedom did not exist, but we should not contemplate the art of the entire history of humanity with the eyes of the 19th century, which is precisely the century in which the figure of the artist began to be valued. as someone who expresses what “dwells in his soul.”

That is, For many centuries, art has been something merely “mechanical”, comparable to the work of other artisans Why not consider what the machines of our time do as art?

We have analyzed how, throughout history, the concept of “art” has changed, and that it was not always the expressive creation, born “from within”, that we still have today.

There were historical moments when art and artists were mere artisans, who we could equate, saving the distance of course, with “machines” that were limited to doing what was asked of them.

However, To what extent can a human being detach themselves from their inner self, their beliefs, their feelings? A person is not a machine, so we will always find in works of art fragments of that internal and genuinely human feeling. Something that a machine, at least for now, cannot do.

Is art, then, that created with artificial intelligence? From my point of view, no. For the simple reason that “art”, despite resisting a specific definition, is closely and inseparably linked to the human being. And a machine is not a human being, and we hope it never will be.