Artificially intelligent systems are slowly taking over tasks previously done by humans, from playing chess to driving our cars. Now an AI system trained on tens of thousands of works of art is producing its own generative abstract artwork, challenging our ideas about what constitutes creativity.
In the movie Her, a man develops a relationship with his operating system, which he thinks of as a person. In reality, however, it’s just a very sophisticated computer program trained on vast databases of real people’s responses and actions. It can process this information and then generate new sentences and actions appropriate to those databases—but, if we are to believe the events of the movie, it has no real understanding of what it’s saying or doing.
It’s possible that artificial intelligence will soon be able to train itself on an even larger database of enriched content: the world’s artistic culture. The goal would be not merely to produce coherent sentences (like in Her), but to create actual art—paintings, songs, sculptures—all generated millions of times faster than humans can by machines that have, in essence, trained for this task their entire life.
The result could potentially be works that are imbued with the essence of human creativity—but not necessarily any of the human baggage that comes along with it: religious belief systems, political and racial biases, social mores—or even awareness that such things exist at all. Granted, many would say that these problems are what make our art so inspiring, but the jury is still out.
Instead, these works could reflect only the deeper principles from which our humanity arises: how we perceive time and space; what we find beautiful; how we express love; how we relate to one another; and how we think about our place in the universe.
This may sound like science fiction (and perhaps it is), but recent advances in artificial intelligence have made this kind of thing plausible—and maybe even inevitable. 1SecondPainting is a great example.