Lance Weiler prepares Columbia University students for the unknown. “What I am about to show you may disturb you.
A filmmaker who made a name for himself at the forefront of entertainment technology by inventing the popular augmented reality game for the movie Head Trauma, Wyler used his experiments for work at art school, where computers We showed you how to be creative. A partner instead of a professional dead end. His classes combine Edgar Allan Poe’s augmented reality, Sherlock Holmes’ virtual reality, and Mary Shelley’s Frankenstein machine learning.
Now, Weiler wants to prepare students for an art world that is slowly embracing the latest digital tools. The widespread availability of artificial intelligence programs that generate images and text are upending the rules of cultural production and changing the way students learn what it means to be an artist.
The 53-year-old professor grinned behind his graying beard. The enthusiasm was like a mad scientist meeting a subject. He spoke to the class in a dramatic whisper.
“This class is about students embracing machines boldly,” Wyler later explained in an interview, shedding his prophetic mystique.
For months he has been working with students for this week’s workshop at New York’s Lincoln Center and this fall’s workshop at Los Angeles’ Music Center where representatives from the arts and entertainment industries are in the audience. I’ve been rehearsing their AI work. I am thinking of hiring young people. Co-created by humans and machines, these immersive performances employ AI programs such as ChatGPT and Midjourney. These AI programs can recreate human creativity by creating scripts and artwork based on algorithms, consuming billions of data points from across the internet. It remains to be seen whether Weiler students will find tools for their own creativity.
“I keep both programs open on screen all the time,” says Behrang Garakani, a 50-year-old aspiring artist who returned to college after a career in video game development. Weiler’s class uses AI to storyboard ideas he didn’t have the technical skills to draw. “It’s now part of my artist’s toolbox,” he said, likening ChatGPT and Midjourney to how Photoshop has become essential for photographers.
During rehearsals, Wyler would often remind his students of something inventor Buckminster Fuller once said. Sometimes he needed to remind himself of it. Although he was no ivory tower expert, he was a self-taught practitioner, and it was with his students that he struggled to understand the emerging technology. His risk of failure and criticism from his colleagues was high.
Haiyu Zhang, a student in Weiler’s class, noted the ongoing debate over the truth of AI-generated ‘pictures’, saying, “Many professors assume ChatGPT to be plagiarism, so all We do not allow students to use ChatGPT in our classes.” “But he makes a point of actually using these tools.”
Zhang, 22, an undergraduate student studying computer science and comparative literature, petitioned for admission to Wyler’s graduate program. Like her other students, she didn’t want to call herself a conventional artist, but she didn’t want to call herself a coder, a futurist, or a worldbuilder. “A lot of people are scared of this technology,” she reasoned. “But not me”
AI: Tool or Threat?
Weiler isn’t the only teacher experimenting with AI art schools across the country, from the California Institute of the Arts to the Rhode Island School of Design. We offer courses for young artists to help them code with machine learning and adopt programs like Midjourney, Stable Diffusion, and more. , and DALL-E 2 — an image generator that turns words into images in seconds. Artists use them as the basis for creative projects that may still involve traditional mediums such as painting and illustration.
In February, New York’s Pratt Institute held a seminar to examine the impact of new technologies. “There’s a lot of precedent for this kind of experimentation,” said Jane South of the school’s art department in an interview. She pointed to previous technological inventions that critics feared would kill the artist’s profession, but only made it stronger. was being held,” she said.
Artists today are expected to create meaning, not just images, and technology can help develop new meanings about modern life, she said.
“The more easily reproducible something is, the more valuable a unique object is to a collector,” says South.
Not everyone agrees. Some working artists (designers, illustrators, animators) view artificial intelligence as an existential threat to their business model.
Last year, a Polish artist named Greg Rutkowski noticed his name being used in Stable Diffusion to inspire fantastical imagery. He argues that the convincing counterfeiting of his own work, created by AI, was confusing buyers. In another case, an artist who works as @ato1004fd on Twitch streamed his drawing session live, but viewers saw his NovelAI using his image generator to complete the illustration. I discovered that Scammers then accused the real artist of being a copycat on social media. And he said in January that three illustrators filed a class action lawsuit against multiple AI companies when the engineers designing the algorithms generated images in the style of living artists, and the image generators were copyrighted. Alleged violations of law and unfair competition laws.
“I withdraw my consent,” Karla Ortiz, one of the illustrators, told The New York Times in February. “That data is my artwork, my life. It feels like my identity.”
A spokesperson for Stability AI, the company behind Stable Diffusion, one of the companies sued, said in a statement, “Anyone who believes this is not fair use does not understand the technology. You misunderstand the law,” he said.
Artificial intelligence research leaders say it’s impossible even for them to understand everything about the technology. Meg Mitchell, chief ethics scientist at AI company Hugging Face, said, “It’s changed fundamentally over the past few years.
Mitchell said many companies now refuse to explain how their programs are getting smarter, citing fierce competition within the industry. For example, when OpenAI released its latest model, GPT-4, the developer declined to share details of the dataset.
“I tend to be chatty about the data,” Mitchell added, suggesting within the industry that the data was “collected without the artist’s consent, without giving credit, without compensation.” increase.
Weiler doesn’t worry about legal challenges to artificial intelligence. “It challenges the status quo of how art is made and who can make it,” he said.
That struggle against a recognized art gatekeeper was once his struggle.
Wyler grew up in a suburb of Philadelphia and joined the local film industry in the 1990s after high school. His early years as a journeyman and camera assistant on large commercial shoots ended in 1998 when he helped direct his indie thriller “The Last Broadcast” on a meager budget of $900. I was. It was hailed as the first fully digitally distributed feature film, grossing over $5 million.
This success prompted Weiler to keep tinkering with new technology. He digitally describes all the gimmicks in his workshop as “fascinating objects,” quoting science fiction writer Arthur C. Clarke.
In 2006, Wyler began incorporating elements of immersive theater and augmented reality games into his films. In this film, the audience interacted with the characters via mobile phones. It’s the kind of history that students keep coming back for more.The Digital Storytelling Lab will be his third course Garakani will take with Wyler. Wailer says, “Extract what you can’t read in textbooks.”
However, he takes a more cautious approach to artificial intelligence than his mentor. “I’m still struggling and inconclusive,” admitted Garakani. He wants companies to be more transparent. “From an academic point of view, please cite your sources.”
Other students suggest that Wyler embraces the magic of artificial intelligence without fully addressing its effects. We can, but we don’t teach students how to build their own algorithms.
Wyler replied that Columbia University offers other courses that go into the computational aspects of artificial intelligence programming. “What we want to do is introduce students to new technologies.”
Tryouts at the New Frontier
In late March, Wyler’s class met inside the Lincoln Center to present their work to an audience of art professionals.
The night started with a provocation. Wailer asked everyone to close their eyes and imagine the values that should be inherited into the future. The students then split into groups and took to the podium to pitch their AI-generated creations.
Zhang and her teammates adopted an ominous tone in “Dream Apocalypse,” an hour-long immersive experience that uses AI to imagine a world where society has collapsed. On the pitch, they described it as “‘Your Worst Nightmare’ meets ‘Inception’ and ‘Russian Doll’ meets Sisyphus.” Audiences are urged to confront their fears by Midjourney’s post-apocalyptic imagery, such as that featuring a child staring into a mushroom cloud of a nuclear explosion.
The Garakani team then shared a mood board, a collage of AI images. This mood board depicts a ruined aquatic civilization staffed by fish-like janitor. Participants joined the goal of cleaning up the forgotten underwater world. A member of the group described the setting as “‘Guardians of the Galaxy’ meets ‘Atlantis’ and ‘Fantasia'”.
However, the concept art seems to riff on “Star Wars” clichés like Baby Yoda and the Death Star, using tools that simply recycle pop culture memes to show how the original artist did it. I was wondering if I could.
These were just prototypes, and the students could then spend weeks refining the text prompts and getting their creativity out of the machine. “Refine, redo, merge, separate, contextualize,” Galakani repeated. “AI needs guidance to create usable jobs.”
Based on feedback from the original audience at Lincoln Center, the Guaracani team switched topics to Carl Sagan, two gramophone records containing the sounds of the Earth that NASA launched into space for aliens to find in 1977. ‘s Voyager Golden Record. Midjourney has created one piece of artwork for the team that looks like a fingerprint. Garakani thought it probably came from a student prompting him with the word “haptic.” This may have triggered the AI program to generate images related to haptic concepts. In other words, the universe in the shape of a fingerprint. The students of the Garakani team were thinking about how participants in their performance could talk to aliens through their golden records.
“We could have gotten to this point without AI, but the machine gave us the idea in seconds,” Garakani added.
More recently, Wyler announced his own AI work in progress, Blockchain Fairy Tales. It depicts a crowdsourced fantasy world using technology related to cryptocurrencies and his NFTs (Non-Fungible Tokens). The artwork will be on display at LA’s Music Center later this year, alongside his students’ immersive experiments.
Zhang said artificial intelligence allowed her to focus on the big concepts while the algorithms filled in the details. She predicted that the wider adoption of these programs would place a greater emphasis on creativity.
“What makes artists special is their ability to imagine something new,” Zhang explains. “So I think AI tools can help us express our creativity, but creativity remains the driving force of the future of art.”
From Weiler’s point of view, his class had no choice but to accept the machine. “What would it look like to slow a cycle moving as fast as artificial intelligence?” he asked. “Well, nobody’s slowing down. You’ve opened Pandora’s box. It’s already unboxed, man.
