Years ago, when the Simpsons animator Chuck Sheetz was a lecturer at UCLA, he invited a producer colleague of his from the series to attend a screening of his students’ work. That was a Saturday night; when Sheetz entered the Simpsons offices the following Monday morning, he found a handful of his pupils in the reception area. The producer had offered them jobs, they said, and they were ready to join the team—the students had the animation skills required of a thriving production.
These days, Sheetz, who’s now a UCLA professor, doesn’t share this anecdote with his students. “Who the hell wants to hear that?” he told me. Even workers in Hollywood today may not want to: The entertainment industry’s post-pandemic years have been marked by a production exodus from Los Angeles, a decline in the number of projects green-lit amid major corporate mergers, and labor strikes. The threat AI poses to both live-action and animated media only adds to these ongoing issues: In a survey published last fall, the executives and workers across Hollywood who responded considered the jobs of animators, visual-effects artists, and concept and storyboard artists among those most likely to be affected by AI-related changes. They’re “just not getting as much work as they have in past years,” Audrey Schomer, an industry analyst and the survey’s author, told me. “If they do have a job, they’re probably being asked to use the tools themselves.”
Although blaming any lack of jobs solely on the rise of generative AI is hard, major Hollywood studios have been gambling that the technology will become a pivotal part of the creative process: In 2024, Lionsgate partnered with the AI start-up Runway, a thus-far-faulty deal meant to create videos off of models trained on the studio’s output. Netflix acquired InterPositive, Ben Affleck’s company that develops AI tools to help with basic filmmaking techniques, in March, and recently began hiring for a division called Inkubator that will experiment with AI-assisted productions. Amazon MGM Studios launched the GenAI Creators’ Fund, an initiative to finance and green-light projects that incorporate the use of AI. And major directors have been using generative AI for previsualization purposes (that is, preproduction design work) and animation, leaving traditional illustrators especially vulnerable.
Amid these developments, some artists in Hollywood are finding the job more and more unsustainable. A few of the ones I spoke with told me that they have received less work in recent years, and one noted that organization tasks once assigned to entry-level workers are now being handled by AI. The industry’s vulnerability, of course, isn’t new; Hollywood tends to cycle through periods of prosperity and contraction, and AI is just one factor in its ongoing changes. Still, “it feels kind of especially rough right now,” Sam Tung, an art director and a member of the Animation Guild, one of the artists’ unions, told me. “There’s a career that I found that I love and I’m really good at, and because of forces completely outside of my control, the viability of it is in major doubt.”
Sheetz, who won an Emmy for his work on The Simpsons, has been an animator long enough to watch technology destabilize his profession before anyone cared about AI. When he joined The Simpsons, in 1991, animation was still a job done largely by hand. Every studio project he knew of involved what he called “an army of cel painters”—artists who carefully inked individual frames on clear acetate sheets placed over light tables.
As computer-generated imagery became the dominant practice, Sheetz remembers other animators around him worrying that the skills they’d developed would become obsolete. Instead, his career took off as the technological advances meant that more films and television shows were being made: “It was, like, a 30-year period of a lot of employment,” he said. But production has seen a recent downturn, which he suggested was a result of post-pandemic turbulence. And though he’s fielding concerns from his students these days about whether Hollywood can survive this period of turmoil, he’s trying to be optimistic about their future. If artists keep up with the industry’s shifting demands, “there’s no reason to believe it couldn’t happen again,” he said, referring back to when he saw his students get hired to work on The Simpsons.
Xindi Zhang, a recent graduate of the University of Southern California’s Expanded Animation Research and Practice program, an MFA curriculum focused on experimental media, indeed found that working with AI paid off. For her thesis film, she trained a program to reimagine preexisting footage she’d gathered and illustrations she’d made; using a technique called “AI morphing,” she rendered those images into watercolorlike clips that evoked the feeling of being in a dream. Last year, Zhang won a Student Academy Award for the project—a prize that put her on the radar of visual-effects studios looking for creative talent familiar with AI software. Before then, she told me, she’d thought about abandoning her dreams of working in Hollywood; now she’s working as an AI specialist at a visual-effects studio.
[Read: The Warner Bros.–Paramount merger isn’t Hollywood’s biggest problem]
Yet the academic departments that once molded students into professional animators are now in danger of disappearing: USC’s Expanded Animation program quietly shut down last month. The California College of the Arts, the state’s oldest private art-and-design school, will be closing in 2027. The reasons for these closures haven’t been attributed to AI’s impact on the industry—more so a mix of dwindling enrollment and funds—but the waning avenues for emerging talent to find a foothold in the business mean fewer chances for such artists to keep up with the industry’s evolution. Zhang considered the Expanded Animation department a “utopia” for her as a student, but, she said, she thinks such animation programs might not be as necessary anymore. “I don’t think you can really teach students the current stuff,” Zhang explained. “There’s no ‘current stuff.’ Every minute is changing.”
Keeping up with “the current stuff” has been as hard on those already working in the industry as it has been for students. In April, Marvel laid off the bulk of its visual-development department, a group consisting of seasoned concept artists who helped turn the studio’s comic-book superheroes into movie stars. One of them, Wesley Burt, received the news in a conference room that featured a mural incorporating art he had drawn of characters on Loki, the Disney+ series—a touch he found ironic.
Lately, he explained to me, many workers like him have seen the gigs they usually take between projects dry up. They tend to draw images from scratch to help directors pitch their ideas to studios, but filmmakers have been leaning on generative-AI tools to do that job. Such technology can create pictures that look appealing but that actually involve nonsensical flourishes that, for example, are impossible to construct into operable sets or develop into wearable costumes. AI, he said, “is not built upon the foundations of people who understand whether it’s set design or understand cloth cutting or how to fabricate, like, a superhero suit or anything like that.” He added, “You need to have people that are highly trained and understand the actual application of it.”
Reid Southen, who began his concept-drawing career coming up with designs for The Hunger Games, has a punchy nickname for concept artists who are being hired to clean up what generative AI got wrong: “slop janitors.” Working closely with filmmakers to go over a script or an outline, establish expectations, and offer ideas for what a character or a world can look like used to be a key part of his job. Now there’s a “slippery slope,” he told me, of artists becoming more familiar with fixing what AI produces than with honing their own artistic abilities. “Your skills atrophy, and your critical thinking atrophies,” he explained.
Although AI is ostensibly being deployed for cost-saving purposes, its inability to understand what’s actually possible in a production can introduce obstacles into a process that used to run more smoothly. “If you’re not an expert in anything and you use generative AI, of course everything is going to look fucking amazing to you,” Jon Lam, a storyboard artist who primarily works in animation and video games, told me. But for those who are trained in more traditional visual-effects modeling tools, he said, “they’re going to be like, ‘Uh, this wheel is connected to a leg; I don’t know what to do here.’”
This ongoing experimentation with AI tools has yielded an era of trial and error—or, as Lam put it to me, “more error than trial”: AI may seem like an attractive means of cutting costs and speeding up the production process, but its messy results leave concept artists less time to come up with original ideas or fine-tune the project. Often, Southen said, he has to work “on autopilot”; he’s concerned about whether he’ll have to meet tighter deadlines for less pay. Like work in many other creative industries, the job has become less of “a process of creation, more of a process of curation,” he explained. Forget coming up with fresh character designs and building singular worlds; for companies using AI this way, much of what’s produced is an iteration of what came before—as in, what got fed into the models themselves.
[Read: What AI will do to art]
AI image generators have already proved troublesome for Hollywood: Last June, Disney and Universal sued the AI firm Midjourney for copyright infringement after Midjourney used images of popular figures such as Darth Vader and Shrek to train its models; Disney also exited its billion-dollar licensing deal with OpenAI earlier this year after the AI company shut down Sora, its video-generation app. (The program would have allowed users to make their own shorts featuring Disney characters.) As much as Netflix has been making moves to incorporate AI, the company has been cautious too, drawing up guidelines last August for the usage of generative-AI tools in its productions. And groups such as the Creators Coalition on AI, which consists of entertainment-industry workers including actors, filmmakers, and executives, have begun establishing the best practices for ethically working with AI companies to ensure that they get permission from artists and compensate them for their labor.
But many artists can’t afford to simply wait and see whether these efforts to protect their livelihoods pay off—and neither, Lam pointed out, should audiences. Technology stealthily changes our viewing habits, he noted: In a span of a few short years, second-screen viewing became a common practice, and the same could go for how tolerant people become of AI-generated work. The creative talent I spoke with commended the public outcry about AI encroachment, including over the Marvel layoffs, AI-generated viral videos, and the prevalence of Super Bowl commercials pushing AI technology this year. But the more the public absorbs such images anyway—and perhaps experiments with making them on their own—the more the definition of quality is likely to blur.
For now, Lam explained, distinguishing what’s been made by AI from what hasn’t been is relatively easy: “If you watch a lot of generative AI, there’s a lot of motion that looks realistic yet somehow feels empty,” he said, citing how an action as simple as opening a door can look lifeless. But over time, perhaps AI will become so prevalent and casually used that it will obscure what makes a character idiosyncratic or a world memorable—the care artists have put into their work (“the little isms,” as he put it).
Take Lam’s latest favorite example: the supremely silly, extremely expressive lizard in the Pixar movie Hoppers, released earlier this year. “You just fall in love with the character by the way they just do something, right?” he said. “And … you can’t really put your finger on it.” And yet, even for workers who’ve been doing this for years, “You’re back to having to prove yourself again,” Southen explained, “to say, ‘Hey, I’m worth hiring.’”
The post A Test Case for AI Creep in Hollywood appeared first on The Atlantic.
