AI-backed robot painting aims to boost artist income
Montreal-based artist Audrey-Eve Goulet was initially uncertain as she watched an AI-powered robotic arm reproduce one of her works, but says that the outcome was “really impressive.”
“I was surprised in a good way,” she says, as she watched the device grab a brush, dip it into a pot of paint, and replicate her work, stroke after meticulous stroke.
Goulet had agreed to work with Acrylic Robotics, a Montreal-based company that says it aims to help artists earn a living by making high-quality replicas of their work, with their consent.
Company founder Chloe Ryan tells AFP the idea began after coming to a discouraging realization about her own income. She said she first starting selling paintings at 14, but grew frustrated at the weeks, or even months, required to make each piece. “I did the back of the napkin math, and I said, ‘Oh my god, I’m making $2 an hour.’”
Ryan studied mechanical robotics at Montreal’s McGill University, and began considering how robots could help reproduce her own work, before launching a company to make the technology accessible to artists worldwide.
The last layer
Assessing the robot’s performance, Goulet says: “It truly looks like one of my works.”
“I like that you can see the strokes… You can really see where the brush went and the shape it drew,” she says, conceding the robotically producing version had “less story behind it” than her own. “My final piece might have gone through five lives before getting to this, but the robot only sees the last layer.”
Ryan says that by replicating “stroke chronology,” her company’s reproductions can capture “the aura of a piece… in a way that a photo print simply never could.”
To reproduce Goulet’s piece, an Acrylic Robotics specialist recreated the work using digital brush strokes and pigments, developing instructions to guide the robot.
Ryan plans to advance the technology, allowing artists to upload images directly. She also wants to create an on-demand market where clients could make special requests, like a portrait of their dog in the style of their favorite artist.
A long waitlist
Ryan says that she understands the artistic community’s concerns about generative AI, but stresses that her company is grounded in the so-called “three Cs” demanded by artists: consent, credit, and compensation.
“A lot of people, before they understand the why of what we’re building, see a robot painting and go, ‘Oh my god, this is the worst thing I’ve ever seen,’” she tells AFP.
But Acrylic Robotics is focused on boosting artist income, especially for those who don’t break into the elite gallery circuit, Ryan says. When approaching an artist, she sometimes suggests they send a few references pieces—work that has already been completed. And when she tells them, “I will just deposit money in your bank account at the end of every month…. There’s a warmer reception,” she says.
The price of reproductions can vary, averaging between a couple hundred to a thousand dollars. And the revenue split with the artist fluctuates. An emerging artist who simply uploads a picture of a piece with limited value may get five percent of a sale, but that figure could rise to 50 percent for a prominent artist with their own base of interested buyers.
“We have a waitlist of about 500 artists,” Ryan says.
Michael Kearns, a computer and information science professor at the University of Pennsylvania, questions whether the technology would ultimately lower the value of the product.
Kearns, part of an Amazon scholarship program that funds academics to work on technological challenges, says he understood the push to “let many more people make a decent living from (art).”
But, he cautions, “When you make something that was scarce abundant, it’ll change people’s perceptions about its value.”
