Among them were Ben Kovalis and Eyal Fisher, co-founders of a new startup called Art AI, which sells unique prints of art generated by neural networks. “We thought we could do much better—we could make tens of thousands of pieces very easily, and each of them would be much better than the piece that sold at Christie’s,” Fisher said.

So the two Israeli tech workers set out to do just that with an AI trained on tens of thousands of pieces of fine art spanning different eras of history. Their online gallery is divided by subject matter—portraits, flowers, nature, abstract and surreal—and filled with slightly garbled images in the style of artists like Monet, Van Gogh and Da Vinci.

The startup is part of a growing AI-generated art scene driven by advances in AI’s ability to mimic human creativity—particularly a type of neural network setup called a GAN (generative adversarial network). AI-generated art galleries and exhibitions have popped up all over the world in recent years, and agencies and brands have begun to explore how this tech might be used for applications such as product design and creative work.

Since the site’s launch earlier this year, Art AI claims to have sold more than 1,200 pieces, which are all $44 each and printed only once per image. The co-founders say their customers are a mix of AI enthusiasts who like the novelty of the concept as well as people who appreciate owning an original work for its own sake.

The company has also begun to commission various human artists, who make use of generative AI tools to produce more polished-looking pieces. While GANs are able to generate wholly original works on their own that resemble an amalgam of the images on which they are trained, Fisher thinks they will ultimately serve more as a tool for artists rather than a replacement for the creative process.

“AI art is going to have a big impact on the art world both as a medium and a tool,” Fisher said. “But I think we will be using AI to enlarge the artist’s world.”