Two of Hollywood’s biggest stars, Tom Cruise and Brad Pitt, engage in a fierce battle atop an abandoned building.
“You killed Jeffrey Epstein, you son of a bitch,” Pitt yelled at Cruise as he blocked multiple blows.
Looks real – too much Authentic. Is it a scene from a big-budget action movie that will be released soon? No, these aren’t the Hollywood stars we’ve come to know and love – they’re just replicas, machine-generated for our viewing pleasure.
Over the past two weeks, Seedance 2.0, a new advanced AI model, has released a flurry of equally bizarre and cinematic videos, causing major concern in the global creative industry.
What is Seadance 2.0?
Seedance 2.0 is a new generative AI (GenAI) model developed by ByteDance, the Chinese tech giant behind TikTok. This is not the first GenAI model. It follows in the footsteps of the likes of Midjourney and OpenAI’s Sora, both of which have the ability to generate video content based on written prompts. However, UTS media arts lecturer Dr Justin Harvey says Seedance 2.0 has attracted attention for its ability to create much higher quality videos than previous “AI slops”.
“The hype around SeaDance seems to be primarily due to its great ability to create believable characters, action and interactions, either between existing humans or between characters like Brad Pitt and the Teenage Mutant Ninja Turtles,” Harvey said.
Seedance 2.0 is currently only officially accessible in China or through a small number of AI creative suites, said Max Schlazer, creative arts subject lead at Swinburne’s Center for Transformative Media Technologies. Regardless, I was impressed with its ability to better render videos of human movement, fast-paced action, and lifelike physics.
“When using GenAI, you typically have to be very smart to maintain visual consistency,” Schleser says. “Before, we would create one video prompt, then another, and then we would have to stitch those different scenes together into a sequence. But with Seedance 2.0, we can do that using visual continuity and multi-shot storytelling.”
The Ultimate Test: Will Smith vs. Spaghetti
One of the reasons Seedance 2.0 stands out so much compared to other GenAI models is because it was the first to pass the “Will Smith eating spaghetti” test.
“[It] “We created a realistic video of Will Smith eating spaghetti,” says Dr Dana McKay, Associate Professor of Interaction Technology and Information at RMIT, “This is a difficult test because we are familiar with Smith’s face and eating behavior, but also because the movements in the scene – Smith, his mouth, and the spaghetti – are all complex.”
How did people react to it?
Although other GenAI models have existed before Seedance 2.0, Schleser says the viral video still made an impact. In one film, a giant Bruce Lee defeats Godzilla. In another scene, Wolverine drives his claws into Superman’s chest.
Entertainment giants around the world quickly took notice of the use of existing famous characters. Major studios including Netflix, Paramount, Warner Bros. Discovery, and Disney have sent cease-and-desist letters to ByteDance, citing copyright infringement.
“Chinese AI service Seadance 2.0 was involved in large-scale misappropriation of U.S. copyrighted material in a single day,” Motion Picture Association President Charles Rivkin said last week.
“By launching a service that operates without meaningful protections against copyright infringement, ByteDance is disregarding established copyright law.”
The American actors union SAG-AFTRA also issued a statement condemning Seadance 2.0’s unauthorized use of its members’ voices and likenesses.
“This is unacceptable and undermines the ability of human talent to earn a living.”
In response, ByteDance said it was taking steps to “strengthen current safety measures.”
Major Hollywood figures have also expressed concern. deadpool & wolverine Co-author Rhett Reese.
“I hate to say it. We’re probably done,” he said. I wrote to X. “Soon, one person will be able to sit down at a computer and make a movie that is indistinguishable from what Hollywood is putting out today. Sure, it’s bad if that person is bad. But if that person has the talent and taste of Christopher Nolan (and such people are rapidly emerging), it will be tremendous.”
What does that mean for the future of film and television?
Some veteran filmmakers are already delving into AI-generated content. Co-author of pulp fiction Roger Avary has partnered with AI Studios to produce three AI-powered films. In other places, black swan Darren Aronofsky’s studio Primordial Soup uses AI to create American Revolutionary War series This day… 1776. And in November, Melbourne’s ACMI hosted the Disrupt AI Film Festival, dedicated to screening AI films.
Indeed, most of these works have been described as “AI slop.” That is until Seedance 2.0. Concerns are now at a fever pitch about how this will affect the creative industries.
Kimberley Weatherall, co-director of the Center for AI, Trust and Governance at the University of Sydney, said SeaDance 2.0 could pose a threat to the film industry’s “middle class”, including storyboard artists, junior VFX editors and Foley artists.
“Will other opportunities arise? Will there be a democratization where some of these people create their own content and have early commercial breakthroughs, or will the opportunities simply disappear? Will there be a second, more ‘artisanal’ market, a ‘human-only’ market?” Weatherall says. “In any case, the concerns are likely to be real.”
mad max Director George Miller said he believes GenAI will make filmmaking more egalitarian. Dr Dana McKay, associate professor of interaction, technology and information at RMIT, said it could create a lower price of admission for people who have creative ideas but don’t have the funds to bring them to fruition.
“Nowadays, it takes a lot of money to make a ‘big budget’ movie,” she says. “Something like Seadance can make it feel like they used expensive stars, great locations, and expensive stunts without using any of that.”
It also has the potential to make filmmaking safer and more time-efficient. Even scenes that required physically dangerous stunts can be generated with AI, McKay says, allowing multiple takes to be combined to create one perfect scene without having to be reshot.
But Weatherall points out that both distribution and marketing – expensive, complex, but necessary processes – are obstacles to AI production.
“Even if you count film festivals, theaters have limited bandwidth with new material,” she says. “I’m worried that expectations will only increase. The ability to produce more, faster, means more is required, faster.”
The cost of AI may also increase. Weatherall currently receives large subsidies from AI companies, but said costs may increase to ultimately turn a profit. “Then we go back to where we started, and only those who can afford computing will produce at that level of quality.”
There’s also the content itself. Because GenAI is based on patterns in data, Weatherall says that as more films are made using the technology, we may see more of them.
“GenAI won’t naturally generate anything truly new or outlier. This isn’t unique to AI. Today’s action movies all seem to have a certain look…but AI will recognize and amplify those patterns and sameness.”
Meanwhile, Harvey said AI could become a complementary tool that helps embellish and enhance human creativity. I did so with brutalistIn , AI is used to refine actors’ Hungarian language, which could potentially be done in action or sci-fi films that require extensive VFX.
He sees these models as tools that can be integrated into production workflows, like an extension of traditional CGI. Therefore, fears that computers will “take over” are largely unfounded. Humans will still need to both drive the AI system and determine its quality.
“I like to think from the audience’s perspective: What is acceptable to the audience?” Harvey says. “I think audiences still want and will continue to want the kind of authenticity that can only be achieved through the human creativity of people thinking, visualizing, drawing, animating, filming, acting, and post-producing things for people.”
Disney recently signed a deal with OpenAI’s Sora to give it access to the House of Mouse’s trademarked characters. In 2024, Lionsgate did something similar with artificial intelligence research company Runway. Schlaser said this type of consolidation will likely continue.
“These deals keep everything within one ecosystem,” he says. “If you have the original source material, it’s much safer than just scraping what’s online.”
Schlaser said GenAI could also be considered a collaborator. We can still use creative human approaches, and AI can help us enhance and implement things that were previously impossible or extremely difficult.
“To understand how a new technology works, you have to work with it, because then you know its limits,” Schlaser says. “You shouldn’t think in black and white” [where] Use all GenAI or none. I think some sort of middle ground will be an innovative approach that looks to the future. ”
Get must-see movies, interviews, and the latest from the world of film delivered to your inbox. Sign up for the Screening Room newsletter.
