The philosophical and existential implications of advanced artificial intelligence models—and the increasing amount of generative AI applications—are complex.
In the coming years, it’s likely that AI will have a major effect on virtually every field. In addition to being used in areas like writing and VFX to perform repetitive work, AI is also creating pieces of high-quality graphic art like film posters or other promotional materials at a faster rate than humans.
In Hollywood, the major concern right now when it comes to studios using AI revolves around replacing workers—a major point of contention in both the Writers Guild of America (WGA) and the SAG-AFTRA strikes. But the major film and TV production companies are reportedly keeping a close eye on AI development in all areas. In May 2023, Disney CEO Bob Iger expressed mixed feelings towards generative AI, weighing “interesting opportunities… and some substantial benefits” for the company against AI’s “highly disruptive” and “difficult to manage” nature, citing concerns over the studio’s intellectual property.
We offer an overview of what AI means for the industry today.
JUMP TO
Many performers are wondering if they’ll one day be replaced by AI.
“The machine learning algorithm doesn’t feel,” explains Hanno Basse, the chief technology officer at Digital Domain, a VFX company whose work can be seen in Marvel Cinematic Universe and “Terminator” films. “And how do you really get the right feeling, the right emotion with the right timing—and also, with the right facial expression and everything else? I think that’s a very human thing.”
Backstage spoke to industry experts to take a deeper look at how AI is currently being used in the acting and casting spaces, plus where the technology might go.
If the influx of memes utilizing the voices of famous people—mainly politicians and celebrities—didn’t give it away, voice simulators have grown in popularity. Currently, generative AI voiceover software like lovo.ai and resemble.ai is impressive—and worrisome, considering how easy it’s become to re-create the voices of public figures.
To learn more about where VO talent and AI are overlapping and what you can do to protect yourself during the casting and hiring process, read the full Backstage Report.
While some companies, such as Levi Strauss & Co., are experimenting with the use of AI models, the application of the technology in the industry is limited. Right now, the main draw for the fashion industry are customizable tools like Lalaland.ai, which let users adjust an AI model’s skin tone, body type, and age. This technology allows customers to see how articles of clothing will look on an approximation of their own bodies.
But as with actors, human beings can express nuance and contribute creatively in a way that AIs can’t. For the immediate future, designers will still need human models to walk the runway at fashion shows.
Influencers looking for an AI-assisted leg up have more tools at their disposal than before. At the moment, generative AI marketing tools can help human influencers who are looking to track current trends and optimize their workflow. Content creators are turning to apps like Notion AI to streamline their campaigns, put out posts faster, and find new ways to reach users.
AI is also connecting brands and influencers. For example, marketing pros can use IMAI to find social media personalities who are the best match for a given campaign.
“Virtual influencers” are also making an appearance. For example, Kuki AI is a chatbot created by Steve Worswick, the senior artificial intelligence designer at ICONIQ, to interact with users on her own site and on social platforms like Facebook and Twitch.
“Kuki is programmed to have an answer for everything,” Liz Snower, the co-founder and chief operating officer of ICONIQ, told Virtual Humans. “To understand the conversation, she uses natural language processing to analyze the information given to her by the user. If you ask her any random question, she will have an answer for you—and chances are it’s going to be impressive.”
AI like Kuki still requires a team of humans to create the content for their social accounts. “I think we are still a pretty long way from having an AI that manages its own social media,” she added.
It’s possible to use tools like ChatGPT to generate essays, articles, marketing copy, and scripts. A recent example is Dramatron, which was developed by the team at DeepMind, Google’s AI research lab, to create AI-generated scripts.
Joe Vanderhelm, the associate director of education for the improv hub Rapid Fire Theatre based in Edmonton, Alberta, had the chance to test out Dramatron for himself. He admits that, though the technology is promising, it’s far from perfect. “It was theater that sometimes felt wonky, or it was theater that sometimes felt heavy-handed, like it maybe didn’t have nuance to it.
“It required a human touch. It gave me something that was, like, 80%,” Vanderhelm continues. “I still had to go through it and massage it a bit here and there.... Dramatron makes a weird lighting call or it makes weird stage directions. [I had to say,] ‘That doesn’t really make any sense; take that out so it flows better.’ I think, at least personally, there will always be the human side to it; someone’s going to have to at least edit.”
In Hollywood, AI has become a contentious subject for the Writers Guild of America (WGA). Prior to the guild going on strike, part of its proposals to the Alliance of Motion Picture and Television Producers (AMPTP) included regulations on the use of AI. These included stipulations like:
- AI cannot write or rewrite literary material.
- AI-generated writing cannot be used as source material.
- AI cannot be trained using material covered by the Minimum Bargaining Agreement.
The AMPTP rejected the guardrails, countering with proposing an annual meeting to discuss advancements in AI technology.
“Mission: Impossible Dead Reckoning - Part One” Courtesy Paramount Pictures and Skydance
New developments in generative AI are allowing filmmakers to put characters into situations that would otherwise be dangerous for actors or their stunt doubles.
Let’s say a director is making a sci-fi film and wants to put their characters in a zero-gravity environment. AI-assisted VFX can make the scene look real in record time. Or what if, in a scene, it’s obvious that a stunt person isn’t the actor they’re doubling? Artificial intelligence can help filmmakers patch the actor’s face into the scene in postproduction.
“The goal is: How can we achieve that in such a way that you can’t really tell the difference between the real actor and the stunt person whose face is being replaced—or a fully digital double?” Basse posits. “And there, we feel like machine learning, especially, can really help to make it look very lifelike.”
Supamotionstock.com/Shutterstock
AI technology has completely changed the game for animators and VFX pros. “Generative artificial intelligence is really [used] to make the jobs of artists easier,” Basse explains. “If you look at animation, for example, there are a lot of intermediate steps that have nothing, really, to do with the storytelling aspects of what an animator does.”
The process of rotoscoping, for example—which involves overlaying animation onto live-action footage—takes hours for animators, as it involves tracing each live-action frame by hand; AI can complete this process in minutes.
“It’s insane how much time it saves; it’s probably 99%,” adds Chris LeDoux, the co-founder and senior visual effects supervisor of Crafty Apes. He says that the VFX industry is experiencing a renaissance thanks to the advent of AI. These new tools allow VFX artists to skip the more menial tasks so they can focus on the creative, intellectually stimulating side of their jobs.
Organizations like AIArtists.org are exploring the possibilities of how artists can work in harmony with artificial intelligence. Graphic artist Julie Wieland, for example, uses Midjourney to create photorealistic imagery.
coup de foudre#aicinema generated in #midjourneyV4, outpainted in #dalle2, color corrected in #photoshop pic.twitter.com/S3CuClK5uj
— Julie W. Design (@juliewdesign_) November 16, 2022
Others, like Refik Anadol, are eschewing realism in favor of vibrant, abstract public art.
View this post on Instagram
“I think what [AI tools] really will do is focus artists more on the creative aspects of things and less on the execution and skill aspects of creating the art,” Basse says. What that means for those in the field depends on what part of the artistic process they most value.
Most artists, however, are less than thrilled about the use of the technology. In December, visual artists began to use the hashtag #artbyhumans on social platforms to take a stand against the incursion of AI and its impact on their livelihoods.
Last year, video game designer Jason M. Allen came under fire when he won the top prize at the Colorado State Fair’s art competition for a piece he created using Midjourney. The scandal sparked a discussion about what the future of illustration might look like, and if the use of AI will cause art to plummet in value.
However, Allen didn’t simply enter prompts into Midjourney and use the first result. He spent about 80 hours finding and editing the best results to create his final product.
PrinceOfLove/Shutterstock
Currently, some musicians are using platforms like WaveAI to assist in the songwriting process, and Google’s MusicLM can help with composing melodies.
Many artists, however, have not been receptive to AI-generated music—particularly as programs get better at creating songs “in the style of” real-life acts.
“What ChatGPT is, in this instance, is replication as travesty,” wrote singer-songwriter Nick Cave, in response to an AI-generated song created in his style. “ChatGPT may be able to write a speech or an essay or a sermon or an obituary but it cannot create a genuine song. It could perhaps in time create a song that is, on the surface, indistinguishable from an original, but it will always be a replication, a kind of burlesque.”