Robots Can Create Art — But Can They Critique It?
WHY YOU SHOULD CARE
Because beauty is in the eye of the (bot) beholder.
By Leslie Nguyen-Okwu
A handsome fellow rolls through the halls of the Quai Branly museum in Paris, snubbing 19th-century tribal art with searing scorn. With a black bowler hat and a chiffon white scarf, Berenson certainly looks the part of a stuffy art connoisseur — so long as you ignore the neural network poking out from his suit. Meet the Art Critic 2.0, built from gleaming metal and sleek sensors, with equal parts smarts and snob.
The sage art critic once commanded considerable power in creative spheres, making or breaking an artist’s career with a simple smirk of disapproval or a punishing review in next day’s paper. But today, as the number of full-time art critics dwindles in newsrooms, a growing force of high-tech art experts is starting to pick up the slack by methodically decoding art’s finest details.
[Robot critics disrupt the] traditional, elite mindset of being told what to think of art, preferably by a white male in a suit.
David Harris Smith, communications professor, McMaster University
In Canada, the Roomba-esque kulturBOT snaps photos at exhibitions and uses an algorithm-powered “stream of consciousness” to tweet out the images with often nonsensical captions like “panting with love of danger” or “streaked with the nocturnal vibration.” KulturBOT then projects these photos and observations onto museum walls for all to see. There’s also the Novice Art Blogger, across the pond in the UK, who mines abstract artworks from the Tate Gallery in London and uses deep learning to classify and conceptualize the images. Translation? Lines of code process the color, texture, line, shape and other elements that make up works of abstract art and then generate impartial critiques. Take, for example, Joseph Mallord William Turner’s 1829 Tynemouth, Northumberland. Imagine “a person walking down the cliff on a mountain slope or rather a picture looking into the sky, on the ground in the water. I’m reminded of a great shot of a mountain near the ocean,” ponders kulturBOT. Carefully phrased with all the pretentiousness of a real art critic, of course.
And let’s not forget Berenson. Named after the late American art critic Bernard Berenson, this robot gleans cues from the facial expressions of its fellow museum goers. With his mechanical mouth, Berenson smiles at art that people generally like and frowns at works that repel others. Granted, Berenson’s quick judgments don’t sound all that nuanced, but unlike most art critics, Berenson and his fellow bots are largely unaffected by personal favoritism, haughty tastes or the whims of a bad mood. Meanwhile, beauty is reduced to pure logic, unsullied by — shudder — subjectivity. The big idea here is to shake up the “traditional, elite mindset of being told what to think of art, preferably by a white male in a suit,” says David Harris Smith, a communications professor at McMaster University in Ontario, and co-creator of kulturBOT.
For good measure, many still doubt whether these hunks of metal and their streams of code can sufficiently internalize the highly complex, intense ideas of art. Yet the history of robot art has a surprisingly long arc, starting in 1973, when artist-engineer Harold Cohen gave life to AARON, one of the first artificially intelligent painters that could produce original art. From standard still lifes of potted plants to 100-foot-long murals, AARON repeatedly passed the Turing Test of the art world, rendering us unable to distinguish between human artwork and AARON’s robo-creations. Since then, more artistically inclined bots have made deeper inroads, including algorithmic artists like Scribble, wall-climbing robot painters like Vertwalker and AI art generators like Google’s DeepDream. The Interactive Robotic Painting Machine in Illinois even responds to sounds in the environment, alters the painting process accordingly. So, don’t be surprised when robots start to curate entire museums or upstage the Mona Lisa one day, says Matt Beane, the chief human-robot interaction officer at Humatics. “Get afraid,” he warns. “Get excited.”
Sure, creativity was once considered the exclusive domain of humans; in fact, few people thought robots could ever be capable of producing such intellect, inspiration and ingenuity with the simple stroke of a brush. However, with newcomers like Berenson and kulturBOT entering the art scene, should critics start scrambling for new careers? Probably not, says robotics consultant Joanne Pransky: “A robot would never get the gut feeling that a human would.” Art requires empathy, she says. And as longtime art watchers all seem to agree on: More important than the superficial visual aspects of art is “the culture, the context, the global trends, the fabrication process — things you don’t see in the final result,” says Daniel Doubrovkine, the chief technology officer at Artsy, the Google of the art world.
But whether or not these robots are good at perceiving art, “there’s something else at work here,” adds Smith. Although robots like Berenson can be as picky about their code as they are with their Cubists, perhaps their function goes far beyond that of an art critic. In fact, “the robot is a kind of artistic installation itself, causing people to ponder why they appreciate an object or not,” says Philippe Gaussier, a robotics engineer and co-creator of Berenson. In that sense, the robots are designed to go Inception on you, forcing you to reflect on what makes art, well, art.
Meanwhile, at a gallery in Ontario, kulturBOT meticulously points its camera at a blank wall and then at a light switch. The resulting caption is anything but poetic: “Standing Futurism, patriotism, the painters and testament to wasted locomotives, alone winter’s nigh.” Fear not, critics, these robots are as ignorant about art as the rest of us.