Putting the art in artificial intelligence
Artists find predicament and promise as they sail into the new world of generative AI.
This story appears in the fall 2023 issue of RED Magazine.
Is artificial intelligence good or bad for artists?
The question has been top of mind since generative AI tools such as ChatGPT and Midjourney were released to the public last year. Tools such as these allow users to input text-based prompts that can generate anything from stylized photographs to romance novels.
Cue the lawsuits. Dozens of artists have filed claims of copyright infringement against generative AI platforms, which use human-created texts and images as training materials. And then there’s the existential dread. Artists want to know, “Will I be replaced by AI?”
While most artists share these concerns, they also recognize that AI is a powerful tool that can benefit their work. Art professors in the College of Letters, Arts and Sciences at Metropolitan State University of Denver are no exception. They realize they must adapt or risk being left behind.
“I understand many people are leery about losing their jobs to artificial intelligence,” said Lisa Ortiz, professor of Journalism and Media Production at MSU Denver. “And while some probably will, AI will also open many doors and create new opportunities. People will not lose their jobs to AI; they will lose their jobs to other people mastering AI.”
Battle royale
Several lawsuits are already underway that place artists and generative AI on opposing sides.
A group of digital artists has filed a suit against Midjourney, Stability AI and the image-sharing platform DeviantArt for using their artwork, among billions of other images online, in training AI. Similarly, authors such as Sarah Silverman and Mona Awad have filed suits against OpenAI, the creator of ChatGPT, for using their books to train AI without compensation or consent.
Copyright infringement may be taking the stage as the most litigious concern, but it doesn’t diminish the fear of obsolescence that accompanies the technology.
“(Generative AI is) something that right now doesn’t have a lot of rules, laws or regulations around it,” said Rebecca Gorman O’Neill, playwright and professor of English at MSU Denver. “Creative writers obviously are concerned because it is a thing that could replace them if it gets good enough.”
Take, for example, the recent strike by the Hollywood Writers Guild of America. Though compensation for writers is at the top of the list of demands, the WGA for the first time has also included a provision for the regulation of “materials produced using artificial intelligence or similar technologies.”
“It’s the first union that is lobbying or demanding that they not be replaced by computers in a creative field,” O’Neill said. “And this is going to set a precedent for a lot of future union action.”
More than 150,000 TV and film actors belonging to the Screen Actors Guild-American Federation of Television and Radio Artists soon followed suit, requesting better pay and guardrails around the use of AI to replicate their likenesses.
The beauty, the magic
For all its drawbacks, generative AI has produced many positive outcomes in the arts.
At the Museum of Modern Art in New York, the installation “Unsupervised” by Refik Anadol uses a machine-learning model to analyze MoMA’s 200-year-old collection and generate interpretations. Natascha Seideneck, associate professor of Art at MSU Denver, who visited the museum last winter, called the installation one of the most beautiful things she had ever seen in her life.
In her own practice, Seideneck has used ChatGPT to help her write artist statements, inputting her current statement and prompting the tool to create an alternative version.
“It wrote the most beautiful artist statement in 30 seconds,” Seideneck said. “And it wasn’t perfect, but there was this paragraph that actually made me look at my work differently.”
Kelly Monico, professor of Art at MSU Denver, echoed Seideneck’s enthusiasm for the technology and plans to use generative AI to help with creative inspiration.
“I’d like to use (generative AI) like a sketch board of 20 different radical ideas,” she said. “And then from there, I use that as an inspiration to create the final iteration of what I’m going to do.”
For Ortiz, the Journalism professor, generative AI allows her to improve her writing. She uses ChatGPT like an assistant or peer reviewer, asking it to read what she has written and make suggestions for improvement. She has also explored Midjourney, creating fun works of art that she would never take the time to create in real life.
“I just think that the new AI tools are fascinating,” she said. “They’re quicker, and quite frankly, Midjourney is a way better artist than I’ll ever be.”
The outer limits
For all the fear and excitement surrounding generative AI, there’s one flaw that can’t be overlooked: The art it creates isn’t all that good. At least not yet.
O’Neill, the English professor, said she can easily spot AI-generated work from her students because it lacks originality and a certain vibrancy, a level of detail that can come only from something AI will never have: sensory and emotional experience.
“It’s not inspired,” she said about the writing generated by platforms such as ChatGPT. In fact, it’s so bad, she uses it as a teaching tool to show students what not to do. “I use it to show the clichés, to show how dead it feels and what adding personal details can do. Your own voice is like a fingerprint, and a computer is not going to be able to copy that.”
4 ways artists can use generative AI1. The launchpad: Generate new ideas and increase your creative output. 2. The editor: Ask for ideas to improve what you’ve already written. 3. The bad example: Analyze its faults as a means of recognizing them in your own work. 4. The experiment: Subvert the tool and find original ways to incorporate it into your process.
|
|
The same is true for image-generating AI, said Monico and Seideneck. The pictures it creates come out a little too digital-looking, they said, and too slick.
“It doesn’t have that humanness to it yet,” Monico said, “but it will.”
For now, artists can enjoy exploring the benefits and learning opportunities that generative AI affords. However, when the day arrives that AI’s artworks parallel those created by people — and most agree that it will — artists will have to find new ways to distinguish themselves.
“I think the key is creativity,” Seideneck said. “Humans just need to be encouraged to push our creativity because that’s what sets us apart from machines.”