Machines can now generate art using existing works as data. But, that begs the question: Are those AI-generated images legit?
In a way, that should come as no surprise. After all, AI-generated art uses historical data to produce work that’s in line with existing expectations. It’s making new art based on established art.
Artificial intelligence uses machine learning and neural networks to understand complex data—data, in this case of AI-art, being images.
Typically, the data is then fed into a system known as a generative adversarial network (GAN), which features an interactive component: One part of the system generates new images (based on the input of old images); the other part of the system judges whether those images are successful—meaning enough in line with the old images.
In other words, AI is trying to make images that fit the established historical precedent. It’s making images that we think are art—not images that it thinks is art.
This concept is perhaps best understood via example: For a piece called “This Sculpture Doesn’t Exist,” artist Matteo Rattini used a GAN to create contemporary art sculptures based on Instagram’s suggestions for popular contemporary art.
Specifically, Rattini created an Instagram account to follow contemporary art accounts, as well as a bot to engage with the art, and then he used the images shown to him via Instagram’s curated list to feed his GAN.
His GAN then created new images, based on the data fed into it, until it ended on “pictures of minimal modernist sculptures showcased inside an aseptic white cube.”
Those images were renderings of sculptures that the GAN determined were art by human standards.
“[AI] reflects how our brain works in a very simple way,” Rattini told Hyperallergic. “You feed it a lot of information on trees, for example, and when it comes time to draw a tree, not only is your idea of a tree informed by all the trees you’ve seen, but while you are drawing it, your brain is constantly judging it through your eyes and giving real-time feedback on how to compensate and adjust.”
Can Machines Replace Artists?
AI is generating new work, but it’s doing so within a carefully derived human context. The work isn’t novel. It’s banal—or worst of all, in the art world—derivative.
“Although AI can technically produce ‘art’ that fits the criteria of a traditional painting (colors, form, and composition), it will not replace the human mind that initially created the source code,” Nina Kong-Suertes, an art legacy advisor, tells Shutterstock.
“AI-generated work is not created from a vacuum, and the result of the work is based on a composite of pre-existing works and conditions that humans initially created.”
Rattini’s sculpture project is a clear example of that. So, too, is Sunspring, a film written entirely using AI.
To create the film, director Oscar Sharp and technologist Ross Goodwin fed the system the text of top sci-fi books and movies. Then, using a recurrent neural network (a machine-learning process similar to GAN), the system wrote a script derived from what it perceived as patterns in other films and movies.
In watching the film, it becomes obvious that the system is merely trafficking in tropes. The script is not new, necessarily, but rather a product of commonalities across top-performing scripts.
For that reason, many people will argue that AI-generated art is not truly art. As Ken Weiner, CTO of GumGum, an AI company, wrote in Scientific American: “Until AI can be programmed to absorb inspiration, crave communication, and want to express something in a creative way, the work it creates on its own simply cannot be considered art without the intention of its human masters.”
But, what if, in the far future, AI does become creative and wants to express itself accordingly? Would that be so bad?
“While many have a dystopian view of AI-generated art as removing the human element of creativity, I believe AI could be an invaluable tool and medium for artists,” Melinda Wang, an art curator and advisor, tells Shutterstock.
“I hope that AI can move beyond impersonation and that artists will harness rich datasets to explore their ideas, use technology to do what can’t be done with paint and canvas, and perhaps create a new 21st century [avant-garde] movement.”
NFTs and Art
The most prolific AI-art movement of the moment involves NFTs, or “non-fungible tokens,” a crypto-currency term for assets that are on a blockchain, containing unique identification codes and metadata which can distinguish them from each other.
PHPs (profile images) are when NFTs are minted as digital images.
These images are largely released in collections—developed from a single concept piece with multiple optional enhancements in the form of hair, eye color, clothing, etc.—and sold for $500 to $10,000 at any given time.
A creator will develop a theme, like Angels, for example, by first establishing a face shape. Then, they’ll develop other items that compliment the face shape—backgrounds, wings, hair color, etc. This results in thousands of images with various amounts of rarity per aspect (i.e. only one or two might have blue hair, or butterfly wings).
Each purchase of an image then gives buyers access to a “community” and allows them to showcase their taste, preference, and values to others. There are 10 to 20 million dollars worth of NFTs trading on the blockchain every week.
This new market hasn’t changed the AI-human teaming process, but has instead leveraged each participant for what it does best. Humans provide creativity and AI enables scale.
We’ll continue to see AI create images derived from and for human pleasure—often with a human involved. These images will more and more closely resemble artwork that we see every day, until we reach a point where AI-generated art will look no different than human-generated art.
And, perhaps in the far far future, we’ll see AI create for its own pleasure, too. That art may end up looking nothing like what we as humans even appreciate, enjoy, or consider art today. Whether that scares or excites you—or both!—it’s worth contemplating.
In the meantime, learn everything you can about NFTs!
Cover image via Login.