A leading creative AI platform is hosting a pop-up shop filled with the work of professional artists who use artificial intelligence tools to create their works, the latest example of a company attempting to monetize a budding machine learning-generated art scene.
Playform AI launched the limited-time exhibition this week with work from four artists, who include established names in the fine art word, new media specialists and a pseudonymous Instagram creator. The pieces range in price from about $60 to $1,500 and include wall prints, jigsaw puzzles, framed video screens and masks. The shop also features an augmented reality tool that allows prospective buyers to project artworks into their home before purchase.
The startup is one of a handful of companies to venture into the AI art exhibition space as new advances in machine learning research have spawned a small but growing community of creatives, artists and technologists attempting to harness the power of AI in art.
The key piece of a technology involved is a form of neural network called a generative adversarial network (GAN), which brands have already experimented with using for everything from deepfaked commercials to product design. Playform has counted brands like T-Mobile, HBO and BMW among the users of its suite of creative AI tools.
The new shop was inspired by a monthly artist-in-residence program that Playform began holding earlier this year. Playform founder Ahmed Elgammal, a Rutgers University computer science professor and director of the school’s AI and Artificial Intelligence Lab, said the AI art aesthetics have grown much more sophisticated in the months since Playform was founded.
“We are beyond the level where AI art has a set of aesthetics that are common between all AI artists,” Elgammal said. “Before, it was all this uncanny look, deformed faces and things like that—it all looked the same. Now, you can see with four artists, four totally different outcomes.”
The evolution has been helped along by platforms like Playform as well as Runway ML and Deep Dream Generator that have made GANs more accessible to artists without the coding experience or expensive computer processing equipment needed to run a neural network locally. Elgammal said that’s transformed the field from primarily technologists with a creative bent to professional and amateur artists looking to the tech as just another tool.
“We are at the point now where it’s becoming a creative tool—it’s an assistant to the artist, it’s under the control of the artist, not the other way around,” Elgammal said.
Jennifer Chang, who handles marketing for Playform, said many of the artists on the platform will train the GAN on images of their own work for hours at a time, checking in periodically to sift through the thousands of visuals it generates for inspiration. The results of the process often look like a somewhat distorted amalgam of the training dataset but can be fine tuned over the course of hundreds of rounds of machine learning training.
Chang said that while the current iteration of the pop-up is a limited-time venture for the holiday gift-giving season, the company is hoping the results will help to give a better sense of the buying audience for a larger marketplace. While the customer base currently consists of primarily traditional contemporary art buyers, Playform is hoping to expand the market with original pieces that are more affordable to the average person.