Jason Allen carried three canvases onto the Colorado State Fairgrounds in Pueblo last August with as much confidence as he could muster. The 39-year-old had never considered himself an artist, but a recent discovery had emboldened him to enter a small art contest: a new artificial intelligence software called Midjourney, capable of creating novel images based on the user’s input. The fair was as ordinary as you can imagine, with carnival rides and fried food, monster trucks and a mullet contest, but Allen’s submissions turned out to be more revolutionary than he expected.
Using AI as his medium, instead of a paintbrush or stylus, he generated a series of images reflecting his interest in fantasy worlds; he also runs a fantasy games company and writes AI-fueled science fiction. He thought of each visual, typed a string of words into a chat box, conjured up an image within seconds and repeated the process — tweaking descriptors — until he was satisfied.
After more than 80 hours and nearly 1,000 iterations, he chose three for his “space opera” series: otherworldly scenes of robed figures communing among the cosmos. The outcome was remarkable. One judge, an art historian, compared them to renaissance paintings. A critic at The Washington Post said they evoke the style of 19th century symbolist painters like Gustave Moreau.
His method could well become a new model of artistic creation. While AI has been developing for decades, its capabilities in the arts have come to a head in recent years, embodied by Midjourney and numerous other art generators launched in the past year or so. There are now cellphone applications that turn selfies into illustrated avatars; text-to-image apps that render photorealism; robots with more of a penchant for the abstract; and chat bots capable of writing anything from haikus to academic papers and even novels.
AI excels at imitation. Trained on terabytes of data, its goal is to mimic human talent and skill — and maybe one day surpass it.
Allen’s submission took first place in the Colorado state fair’s amateur “digitally manipulated photography” category and made national news several times over. The concept wasn’t exactly new; world class institutions now proudly display AI art. But as an outsider, Allen’s win ignited controversy. Across social media, artists have split into factions: those who decry AI art as a form of cheating, a sign of laziness, a harbinger of “literal hell on earth” versus those who defend it as evolution in style and accessibility, or at least a useful tool.
In Allen’s view, the question is utilitarian. “I think art is just bringing vision into form, and I believe that using artificial intelligence tools accomplish that,” Allen says. “There’s something inside of us that’s happening. And then we want to tell other people about it or show other people what that thought or feeling or expression is. We’re all people. And we’re all imaginative and we’re all creative and we all want to express ourselves.” But in a larger sense, his work is a window into deeper questions about the nature of art and humanity itself.
Last fall, AI made its debut at the Museum of Modern Art in New York City, a few blocks south of Central Park. In “Unsupervised,” an exhibit by a Refik Anadol, a video installation flashed floor-to-ceiling murals of dreamscapes: soft beige ripples akin to curves on marble, fibrous lines that look like pencil strokes strewn to connect amorphous shapes. A machine created these dizzying distortions by sifting through data of the museum’s own archived collections. It reimagined art history just four floors beneath pieces that exist to preserve it — like Vincent van Gogh’s “The Starry Night” and Salvador Dalí’s “The Persistence of Memory.”
If there is one thing AI excels at, it’s imitation. Like humans, AI learns through neural networks. But where people learn from experience, AI generators are fed and trained on terabytes of data. They pull from what humans have created and shared online, then analyze patterns to make associations and predictions. The goal is to mimic human talent and skill — and maybe one day surpass it. But the tool is still an amalgamation of human creativity.
Imitation may sound like a pejorative term, but it has been part of how we define art since the age of ancient Greek philosophers like Plato and Aristotle. Both defined art as imitation, but Plato believed that imitation to be superficial, a cheap dilution of the natural world — itself an inferior expression of the ideal — some may say divine. On the other hand, Aristotle — Plato’s former student, a voice of a younger generation — saw art as a way to pay reverence to the temporal lives we live. A way to make sense of the experience and hold onto it for longer: “The aim of art is to represent not the outward appearance of things, but their inward significance.”
Plato and Aristotle also saw creativity as a gift from the gods, the result of divine inspiration. And there is something deeply human — some may say spiritual — in the drive to create, to craft something, to bring intention to life. The first cave paintings some 40,000 years ago are believed to have been made for symbolic or religious functions. From Egyptians to Romans, from Buddhists to Christians, humans have built and sculpted and painted to express reverence for the sublime and curiosity for what was possible beyond their ordinary lives.
AI’s methods of emulation are more advanced than marble statues and terracotta pots. But if art is a tool we use to commune with the divine and make sense of our lives, can algorithms perform the same function? Can a computer grapple with mortality, or touch the human heart? To seek the same results while bypassing the human component seems to risk cutting out the reason we create at all.
Hayao Miyazaki, the mind behind Studio Ghibli’s internationally acclaimed animated films like “Spirited Away” and “Howl’s Moving Castle,” doesn’t think so. “I strongly feel that this is an insult to life itself,” he said, speaking of AI-generated animation in a documentary about his work. “We humans are losing faith in ourselves.”
Guillermo del Toro, an Oscar-winning filmmaker who most recently directed a stop-motion version of “Pinocchio” that won the Golden Globe for best animated feature, echoed Miyazaki in a December interview. “I consume and love art made by humans,” he said. “I am completely moved by that. And I am not interested in illustrations made by machines and the extrapolation of information.”
Even Elon Musk once suggested artificial intelligence was like “summoning the demon.”
Anadol disagrees. “For me, art reflects humanity’s capacity for imagination. And if I push my compass to the edge of imagination, I find myself well connected with the machines, with the archives, with knowledge, and the collective memories of humanity,” he said in an interview for MoMA. “There’s a collaboration between machine and human. With the same data, we can generate infinite versions of the same sculpture, but choosing this moment, and creating this moment in time and space, is the moment of creation.”
On a more practical level, artists, writers and other creators around the world are struggling to determine what it means for them.
ChatGPT, a popular AI text generator, has mastered human-sounding speech. Emily Silverman, creator of the California-based medical storytelling community The Nocturnists, found the chatbot could write something as intimate as a haiku about a colonoscopy:
The cold, sterile room
A long, flexible tube explores
My insides on display
“There’s a vulnerability to that last phrase,” she says. “Maybe it knows that a colonoscopy can be an embarrassing thing for people. Or maybe it knows that, generally, when people get sick they’re in a hospital gown and feeling kind of vulnerable. Maybe it played into that.”
When I took the experiment a step further, asking ChatGPT to analyze its own poem, the program echoed Silverman’s conclusions. “The last line ‘My insides on display’ highlights the vulnerability of the patient during the procedure,” it wrote. “The procedure is invasive and the patient is exposed both physically and emotionally as the doctor examines the inside of their body. The haiku uses imagery to convey the feelings of vulnerability and exposure that may come with having a colonoscopy.”
Silverman admits that the technology is impressive, but she still fears how it might impact the way humans create and learn, especially in a society where few readers take the time to explore poetry or literature for deeper meaning. Searching text, she believes, should be an exercise in discovery. Using AI is a shortcut that could miss out on that. “There’s pain and difficulty in bringing a story into creation that torments me,” she says. “And to think that a robot could do it, that just felt threatening to me.”
Plato and Aristotle saw creativity as a gift from the gods. And there is something deeply human — some may say spiritual — in the drive to create, to craft something, to bring intention to life.
As a buzzword, AI is often associated with that kind of panic, especially among people whose jobs it could threaten. But the technology has long hidden in plain sight — often operating innocuously. Similar technology is at work in our manufacturing plants, our cellphone’s facial recognition, our email text predictions and social media advertisements. Proponents have long preached that it can boost industry. Others are now suggesting its prowess in imitation might even be able to take the arts and humanities as we know them to new levels.
Elizabeth Callaway, an assistant professor of English at the University of Utah, received an award last year from the National Humanities Center to develop a class on responsible AI. She is part of a cohort representing 15 universities around the country, and among the optimists who maintain their faith in humanity when it comes to the AI art conundrum.
Her incoming class on AI will incorporate lessons with ChatGPT where students critique the program’s prose, advancing the next generation of human writers in tandem with technology. Callaway doesn’t believe AI art will replace human creations, but she does believe human art can grow because of it. She’s begun imagining new lesson plans: students grading essays generated by AI, fielding analyses of writing style with side-by-side comparisons of famous authors.
“I think what’s really cool about it is that it is a tool for playing,” she says. Callaway gave an example of asking for two sonnets from ChatGPT, one in the style of Shakespeare and the other in the style of John Keats. “You can then have a fantastic discussion about where it succeeds, whether the two sides are different. Students can figure out how language works with it.”
The same could hold true for visual art. AI can be a hindrance, or it could become a learning tool. Regardless of utility, it’s already changing the artistic landscape. That has happened before. Machines completely transformed the textile industry in the industrial revolution. Photography became a faster and more affordable alternative to portraiture in the 19th century. Buttons and levers in lieu of human hands. Yet centuries later, people still sew and sit for paintings.
Technology is bound to bleed more into the art world regardless of public outcry. The MoMA, for one, already plans to expand its acquisition of digital art. Exhibits like Anadol’s could soon become more frequent, from taboo to norm.
I missed Anadol’s installation, but as I looked over the images online, my eyes were drawn to an etching. Or what looked like an etching. Smears of scarlet blobs screamed against a bright blue background. The colors and shapes felt distinctly separate, worlds away from one another, yet they were tethered together with strands thin enough to mimic spider’s silk. The more I looked the more I thought of that silk, how dainty it seems for a substance that’s stronger than steel. How something can appear impossible to construct then turn out sturdier than you could ever imagine.
This story appears in the March issue of Deseret Magazine. Learn more about how to subscribe.