Would you pay $400,000 for a painting made by a computer?
AI-generated artwork challenges what is considered real art.
Our editorial transparency tool uses blockchain technology to permanently log all changes made to official releases after publication. However, this post is not an official release and therefore not tracked. Visit our learn more for more information.
Briana Brownell has long been a programmer and an artist, but never expected the two pursuits would connect, at least not in the way they have. Her day job, as the founder of the Canadian data science company Pure Strategy, tends to limit her creative side. But her time away from work allows her to experiment and create using artificial intelligence (AI).
“Given my career in data science, I figured I could turn some of the typical tools I use into something unique when it comes to art,” she said. “It’s really exciting to think there can be a whole new way to create art.”
Brownell is part of a new generation of artists using AI to create work. Tech advances have made it easier than ever to create high-quality images with a computer.
And the art world is taking notice. “Edmond de Belamy,” a portrait of a fictional aristocrat developed by the French collective Obvious, sold for more than $430,000 at a 2018 Christie’s auction, raising interesting questions about the value and worth of computer-generated art.
How exactly does a computer “make” artwork? Generative adversarial networks (GANs) and other algorithms allow creators to build a stream of images based on a concept or a base picture. Brownell, for example, uses colors and textures as the guidelines for her AI art, which include trees, landscapes and even album covers. Other artists reshape paintings from the past with the new technology.
Sacramento-based artist Mikko Lautamo uses AI to create generative abstract animations that focus on the methods and movements of complex systems. His work, which often features geometric shapes and lines, are best translated in video form. (Earlier this year, Lautamo took part in Verizon’s Engineers & Artists pilot program, in which participants in those fields come together to create unique artwork based on a theme inspired by 5G.)
The creators of “Edmond de Belamy” utilized data from more than 10,000 portraits dating from the 14th to 20th centuries, then used two algorithms – one that attempts to make a new image based on the set, and another that tries to spot the difference between man-made and computer-generated art to create a “new” image.
Okay, “Edmond de Belamy” proves an algorithm can create valuable art, but can a computer create important, emotive work?
“It’s always a tricky question to determine what is really art,” said Luba Elliott, a UK-based curator, artist and researcher specializing in artificial intelligence in the creative industries. In her role, she’s attempting to find the right mix of art and science.
“When I review art, it’s based on two things: What idea does the artist have? What is he or she trying to do?” Elliott said. “And then, there’s the technological execution. Are they using a sophisticated algorithm? Are they producing a state-of-the-art result?”
Luba noted that people with a strong tech background are developing algorithms that can produce beautiful images, but the concept is “not as developed.”
Elliott hosts events for AI artists across Europe, where she links the tech and art communities as they explore this new territory. She also curates the AI Art Gallery, which hosts a selection of works that use machine learning.
Some critics say AI-generated art isn’t “real,” and while there are some limitations in the creative process, Elliott argues that there is still a human touch in the final product.
“I can understand how some might think the human involvement or scale might be less obvious since it’s combined with the machine,” she said. “It’s very difficult to compare to something that has been purely made by a human. Ultimately, I think there’s a lot of human involvement. They decide to use AI, pick the data set, choose the appropriate algorithm, tweak the design and curate the output to create some kind of narrative to display the image.”
With the proliferation of machine learning, AI-generated art will likely expand far beyond the galleries.
“There are a lot of possibilities with where it could go,” said Canadian AI artist Brownell. “Outside the gallery circuit, the availability for everyday people to use it and express themselves is exciting. They might be able to take some of these tools and have a personal experience creating art for themselves. That’s just as important as art that you share with other people.”
For more information, see:
For related media inquiries, please contact firstname.lastname@example.org