Picasso, Van Gogh, AI? Since artificial intelligence (AI) seeks to mimic human intelligence, it’s natural that AI developers would apply the technology to an activity that is uniquely human: art. Still, while AI is now creating art, it isn’t replacing humans. It’s augmenting us.
Below, I’ll outline 3 impactful ways stakeholders across the art industry are utilizing AI to help artists and art lovers alike.
1. Creating art.
With all the complexities about what makes art “art,” there’s some debate on whether AI is actually creating artwork. Nevertheless, in the last few years, researchers have used AI for various art projects.
The annual RobotArt competition challenges its entrants to “to produce something visually beautiful with robotics.” Launched in 2016 by Andrew Conru, he saw it as an opportunity to combine his two loves: technology and art and encourage others to do the same. Since then, teams from around the world have built robots, some operated by neural networks, that can paint quality portraits, landscapes, and even abstract works.
Fans of the popular fantasy series “A Song of Ice and Fire,” on which the tv show Game of Thrones is based, have long waited for its penultimate volume “The Winds of Winter.” One such fan trained an algorithm to use the information previous volumes in the series to write its own version of the anticipated novel. Though the chapters are riddled with grammatical and plot errors, the AI was able to copy the author’s grammatical tics and produce 5 full chapters. From the first chapter:
Still, this won’t ever truly replicate the magic created by the series’ human author. So, while AI has been applied to art, it’s unlikely to match the skills of human artists any time soon. Even then, some suggest it will still be unable to replace us. From Conru: “Human-generated art will always be highly respected not only for its creativity but in our shared human experience. While we may be impressed by AI chess software, we are thrilled and impressed by human grandmasters.”
2. Learning about artists.
Many of history’s greatest artists produced thousands of art pieces across various forms, but most of us are unlikely to see them all. With that, we are perhaps limited in our knowledge of what captured the imaginations of these men and women. Using AI, we can now analyze these myriads of artworks more than ever before.
Using Clarifai’s computer vision technology, the team at Artnome, one of the largest analytical databases for artwork, built a model to analyze Edvard Munch’s 1800+ paintings. Using several predefined concepts like trees and snow selected from their research, the model was able to quickly identify and tag concepts in Munch’s paintings. This allowed the team to easily curate the artist’s works by subjects.
For instance, going based on Munch’s titles, the team identified 104 paintings that included portrayals of “women” or “girls.” With AI, however, the team found over 500 of Munch’s paintings depicted women and expected that number to increase as the model became more accurate.
Their founder, Jason Bailey, expects AI to open up new possibilities for research. AI, for instance, would be able to tell us how often Munch depicted certain subjects, like fruit or even specific types of fruits, during his career. It may be able to tell us whether the frequency of these subjects shifted as he aged. So, while computers cannot replace artists, using their capabilities, we can gain an even high level of understanding of the brilliant human minds behind the brushes and possibly learn the “ why” behind the innately human need to create art.
3. Selling art.
Online retailers are already using computer vision to elevate the shopping experience for their customers and those that specialize in artwork, like Artfinder, are following suit. With AI, Artfinder was able to develop “Emma,” a Twitter bot that helps customers find artwork that is similar to ones they like. All customers do is take a picture of something that they like, like a sunset, log on to their Twitter accounts, upload the picture and tweet it to Emma’s handle. The bot will then respond with examples of original artwork within Artfinder’s portfolio.
Hey @art4you_art4me, I've found these artworks inspired by your tweet! Head to @artfinder to see more » https://t.co/dAguek4gtJ pic.twitter.com/3lQ1EW1gwQ
— Emma @ Artfinder (@ArtfinderEmma) August 31, 2018
These snap and search functionalities can lead to increased revenue, as they allow customers to quickly and easily find and buy artwork that is specific to their tastes. Where it is part of their product offering, customers can even find decor that complements these art pieces, so we can expect other companies in the space to take full advantage of this technology in the coming years.
With all its mechanical terms and definitions, AI and art may seem to be complete opposites. However, as shown above, the two can be very complementary. Though developers are teaching AI to match our artistic abilities, there is no reason to think it can replace us. Much like a pen or paintbrush, AI is just another tool borne from human creativity to help us to keep up with our limitless imagination.