Artists embrace, reject benefits of AI

Proponents say it’s a tool, not a replacement for the creative spirit.
"How I Will Retire" is an AI generated artwork by Alex Kvares, formerly of Atlanta but now based in New York. 
(Courtesy of Alex Kvares)

Credit: Alex Kvares

Credit: Alex Kvares

"How I Will Retire" is an AI generated artwork by Alex Kvares, formerly of Atlanta but now based in New York. (Courtesy of Alex Kvares)

At the 2022 Colorado State Fair, an AI generated work won a top prize in the digital art category of the annual art competition. The image of a vast, ornate room populated with people in historical garb and dominated by an enormous round window into the universe was created by Jason M. Allen using the Midjourney image generator.

And in May a bestselling fantasy novel “House of Earth and Blood” by Sarah J. Maas published by Bloomsbury incited controversy for the AI generated image of a wolf’s head on the book’s UK cover. Some have speculated that the design was a cost-saving measure. The choice led to concern over how book cover artists may end up as one casualty in the AI revolution.

In both cases outrage ensued, much of it voiced by artists worried that their livelihood could be affected by the wide scale adoption of AI image generation.

As the world ponders the future role and potential impact of artificial intelligence, visual artists are engaged in their own questions and concerns about how AI will change their work. Artists tend to be innovators and early adopters when new technologies emerge, and so many of them have been experimenting with AI.

This is not the first time the art world has grappled with issues of authenticity that arise when a new technology enters the fray. The invention of the camera in 1816 inspired its own doomsayers who considered a photograph an inferior artistic object. Many skeptics believed the camera could only capture reality and not subjectively render it, as a painter would.

Decades later, Photoshop, color photography and digital cameras had their own skeptics among art world purists.

An image generated using artificial intelligence from Austell artist Evan Jones.
(Courtesy of Evan Jones)

Credit: Evan Jones

icon to expand image

Credit: Evan Jones

“People blow the implications of a new technology or tool out of proportion,” noted Austell-based artist Evan Jones. “I guess sculptors are out of a job now that 3D printers have been invented? Hardly. Painters are still around even though there are cameras.”

In many ways, AI image generators like Stable Diffusion, Midjourney and DALLE-E 2 represent the astounding possibilities of a technology that recasts a world of our own creation in startling new form.

By feeding text prompts into those AI image generators, artists can pull from billions of online images. The results are strange collisions of bodies, landscapes, facial features and reshaped, reimagined images of our physical world. These AI-generated images feel like the internet’s subconscious bubbling up a strange, transmogrified new reality.

Alex Kvares, formerly of Atlanta but now based in New York, likens the process of feeding text into AI image generators to the dopamine-rush of gambling. Using AI image generators is, he says, “weirdly addictive.”

Working from home, or even on the subway, Kvares feeds prompts like “muppets” or “transparent flesh” and the names of character actors like Amanda Plummer and Wallace Shawn into his preferred AI tool Stable Diffusion XL. Then he watches the uncanny, often shocking images that are returned. His haiku-like strings of words generate genuinely unsettling visuals where the human and the animal blend, where mayhem reigns and sinew and guts are worn outside the body instead of beneath the skin.

In many ways, the disturbing, anarchic images look very much like the delicate, finely wrought, anarchic drawings that Kvares did before he began experimenting with AI.

“My goal is as much to confuse the AI as to make something I want,” said Kvares of his methodology. “That’s where it gets interesting.”

He will exhibit his AI work at Art Basel Miami Beach in December and in an online exhibition in January on the mepaintsme.com website. The AI images he has shared on Instagram have generated conversation among fellow artists, as well as moments of skepticism. One Instagram viewer commented: “Sell. Out.”

For Kvares, there is no real moral dilemma or controversial take on his use of AI in his art making.

Instead, his challenge is translating what he is making into a marketable object that he can sell at a gallery or art fair. His brave new world is adapting his work to the art marketplace.

Some artists have expressed fear that AI generators will allow their work to be sampled without their consent. Kvares doesn’t buy it.

“People complain about someone stealing their style or sensibility, but there’s nothing new about it. You can do the same thing with a paintbrush,” said Kvares.

An image generated using artificial intelligence from Austell artist Evan Jones.
(Courtesy of Evan Jones)

Credit: Evan Jones

icon to expand image

Credit: Evan Jones

Austell artist Evan Jones agrees. From his point of view, there’s really no need for debate among artists about the advantages or disadvantages of using AI.

“If you’re looking at it objectively, you won’t be on a side,” he said. “AI is simply a tool. The AI doesn’t think, the artist does.”

Jones has dabbled in using AI as a jumping off point for his paintings, which often comment upon Southern stereotypes. He has used AI prompts around “Appalachia” to summon up visions of the region where he grew up.

The results hover between absurd and unnerving. In one image a burly bearded man in a Carhartt jacket confronts an enormous panther against a snowy landscape. On the plaintive end of the spectrum, a weathered looking old-timer kneels in prayer, his faithful dog gazing heavenward in soulful solidarity. Religious ecstasies and feats of derring-do abound in these works that feed into certain visual stereotypes about the rural South. As many have pointed out, AI is created by humans and so it comes with data sets that reflect human prejudices.

"Headshot (Heads series)," (2021) by Mark Leibert was inspired by a drawing created using GAN (generative adversarial network), a precursor to today's artificial intelligence bots.
(Courtesy of Mark Leibert)

Credit: Wyatt Kane Wyatt@wrk.studio

icon to expand image

Credit: Wyatt Kane Wyatt@wrk.studio

Many artists use the analogy of sketching when describing AI’s usefulness. In 2019, Georgia Tech professor and artist Mark Leibert worked with a precursor to AI, GAN (generative adversarial network), to generate images he later turned into paintings. There was value in that interaction, he says, because it helped him see the importance of the artist’s hand in artmaking over the magic of an image created with AI.

It prompted him to ask the question, “What makes a human’s work unique from a machine’s work?” For him, what an artist creates by hand still feels important and unmatched.

“It’s layered, and it’s personal, and it’s emotional — all those things.”

Jones has reached a similar conclusion. His experiments with AI in creating a series of Appalachian portraits delivered unsatisfactory results. Rather than inspiring him to make paintings from those images, they created a very different effect.

“There was no life in them,” he said. He never bothered to create paintings based on those AI images.

Whether artists consider AI a creative tool and inspiration or a poor replacement for artistic intent and execution, there is no doubt that AI image generation tools, like the internet or cell phones, have forever changed our relationship with image-making.

“The way I see it is, there’s no way back from it. It’s not going anywhere,” said Kvares.


ART EVENT

Mark Leibert. Open studio for Atlanta Art Week. 1-6 p.m. Oct. 7. Murphy Rail Studios, 1870 Murphy Ave. SW, Atlanta, Studio 12. www.markleibert.com