‘Pondering’ by Vincent Favilla, created using the Midjourney AI image generator. (Vincent Favilla/Midjourney)
Editor’snote:This story is part of KQED’s Youth Takeover. Throughout the week of April 24-28, we’re publishing content by high school students from all over the Bay Area.
or most of human history, art was a skill that took years to perfect. Now, it only takes a few words and a click to create a masterpiece.
I’m talking, of course, about AI-generated art — a topic that has been the subject of endless debate on social media over the past few months. Every day when I scroll through my phone, I see some people sharing their beautiful creations, while others protest against the ethics of this technology.
For high school students, like me, who are interested in pursuing a career in art, the idea of AI art can be especially frightening. As it becomes easier to create high-quality artwork using this technology, what will happen to the artists who were once hired for these tasks? Why would companies commission an artist when they can create a beautiful illustration or design with just a few clicks? What will happen if the future of art gets completely taken over by technology?
Mo Koelle is an artist and student at Milpitas High School who does digital illustration and hand-drawn art. They feel that AI art generators steal from human artists. “It’s theft,” they say, “using the art that someone has spent hours of their life to make in order to spit something out in a matter of seconds.”
And Tyler Tran, a Milpitas High School senior who’s planning to major in digital art and animation, is concerned that the popularity of AI art might make his entire career path obsolete, especially when factoring in how expensive an art education tends to be. He worries, “Commissions, digital work — suddenly, of all that could be done with just AI. What am I supposed to do? [I need to] pay for expenses, pay for my loans, pay for life, pay for my college, and stuff like that. So that’s definitely a fear that I have.”
To see for myself what these AI generators are capable of, I decided to try creating my own AI art. The first time I did, I was blown away by how powerful the technology was. To create an image using the most popular AI image generator, Midjourney, all you have to do is type in a text prompt — my prompt was “anime-style portrait of girl with brown hair and gray eyes against a forest backdrop.”Just a few minutes later, the AI generated four different portraits.
I was stunned at how good the art was. One portrait had such cinematic lighting, it looked like it could have been from an animated film. Another was more stylized, similar to a still image from a fantasy RPG video game cutscene. Yet another was a more realistic portrait, reminiscent of the kind of illustration you might find on the cover of a novel.
The AI-generated images were all surprisingly detailed. Looking closely, you could see the veins of the leaves, the reflection in the characters’ eyes and the definition on their neck and collarbone. The AI showed a skilled knowledge of light, shadow and proportions.
A single one of these images might take a human artist hours to draw, color and render. The Midjourney AI, on the other hand, completed multiple images in less than two minutes.
This ease and speed of AI image generators have led many companies to use them instead of hiring human artists. AI art is being used to create book covers, album art and even music videos. For example, both the novel cover for Fractal Noise by Christopher Paolini and the music video for “Make Me Feel” by The Chainsmokers were made using AI. As a result, artists — and young people who aspire to become artists — are becoming more and more worried about the future of their profession.
To understand the controversy, it’s helpful to first understand how AI technology works. Most AI image generators function by “learning” from a dataset compiled of countless images across the internet. Vincent Favilla is a psychology professor at Skyline College who is also an artist and AI developer knowledgeable in ChatGPT, GPT-4 and other AI technologies. According to Favilla, AI image generators use “a complex system of matrices and mathematical equations,” known as a neural network, which figures out the relationships between the billions of images and captions that it contains in its dataset. After this, the AI model fuses this knowledge together to generate an image.
Since the AI datasets pull from images online, they also learn from the work of real human artists and become capable of accurately replicating their styles. This has caused quite a backlash. Many artists consider it theft when AI image generators can copy their art style without their consent.
For Favilla, however, it’s an oversimplification to say that AI image generators are stealing images from artists. Since the AI dataset contains billions of images, any single individual’s artwork might make up just a minuscule fraction of that dataset. As of now, Favilla says, it is virtually impossible to use an AI generator to trace and retrieve a specific image. And most AI generators are not fully able to replicate a specific artist’s style.
Koelle, the aforementioned student artist, is one of those who worry about protecting the rights of artists. And though they are not completely against the technology, they ultimately feel that AI generators are “not so much a tool for inspiration as it is a tool for making a product.” While AI has the potential to help artists, Koelle says, “There’s a line between using AI to assist you and to do all the work for you.” And while these AI generators are able to create beautiful pieces, they still rely on humans to input an idea. They aren’t fully capable of coming up with their own unique ideas with personal meaning behind them.
AI image generators have other limitations as well. For instance, they struggle with drawing hands and complex human anatomy and with writing coherent text. Sometimes the results create an uncanny valley effect.
Misha Chaturabul, a Milpitas High School student with a passion for drawing, painting and digital art, says she would like to do art as a side job in the future. According to Chaturabul, “Real artists’ work is made with a genuine passion and love that cannot be replicated by a computer or a type of technology.” She believes that the time and effort that a human artist puts into each piece gives it a special quality — a human spirit or essence that’s able to evoke emotions — which AI cannot reproduce.
Tran also believes art is a uniquely human way of communicating and expressing feelings. “While AI art can have a vague sense of this, there’s something about the personal way that individual artists give a piece of themselves whenever they’re making art,” he explains. “I think we’re kind of jumping from the process, because the process is just as important as the end product.”
For him, the idea of having to compete with AI in addition to other artists feels especially brutal. “We’re really approaching this very modernized, ‘oh, everything must be efficient’ route,” he says. “And I think that’s so terrible, because it’s really just cutting corners. It’s only profiting big companies.”
Tran’s hope is that AI doesn’t completely take over the art industry in the future. “I think there is a point where AI art has its limits, and I hope its limits stay there,” he says.
People like Favilla, on the other hand, have a more optimistic view of the technology.
“I have more ideas than I do time,” the professor says. AI art allows him to realize his ideas more quickly than he would be able to do using traditional illustration methods. The technology can generate ideas that artists can use as inspiration for their own pieces. Moreover, because the AI tools are so accessible, they allow people who don’t have a lot of artistic talent or physical strength to still be able to create beautiful artwork. All you need is a good imagination and creativity.
In other words, maybe AI software is just another medium that artists can use to express themselves.
I have my own concerns about how AI might affect my future. Since I was young, I’ve always wanted to pursue some sort of creative career, such as art or design. Right now I’m thinking of becoming a UX designer. But recently, I saw how the artificial intelligence program GPT-4 was able to turn even just a rudimentary napkin sketch into a fully operational website. That made me feel very worried about whether I’ll still be able to have a career.
Nevertheless, whether we like it or not, artificial intelligence will continue to be part of our lives. I believe that these image generators do have the potential to help people as long as some restrictions are set to protect artists. And, like so many of the aspiring artists that I spoke to, I think there is still something special about being human that AI will never replace.
“Be hopeful,” Tran says. “Don’t feel defeated because AI art simply exists.” After all, he says, the human mind and body are always taking in experiences from actually living in the real world. “That’s something that will always set you apart from AI art.”
Kaitlyn Nguyen is a junior at Milpitas High School who enjoys art, design and everything creative. When she isn’t contemplating life, the universe and existence, she’s reading her favorite novels or making jewelry and designs for The Starlit Shoppe.
Care about what’s happening in Bay Area arts? Stay informed with one email every other week—right to your inbox.