After a difficult year marked by two bitter strikes, the 96th Academy Awards will give Hollywood a chance to project an image of itself exactly the way it wants to be seen.
This year’s nominees include critically acclaimed blockbusters (Oppenheimer and Barbie), edgy fare (Poor Things), auteur turns (Killers of the Flower Moon, The Zone of Interest) and biting satire (American Fiction). The “Barbenheimer” Oscars should quiet recent criticisms that the nominees were too obscure, at least this year.
The line-up is rich enough for the industry to forget, for one night anyway, about the belt-tightening under way at entertainment companies that has made “survive to 2025” a motto for some Hollywood workers.
But a new worry has emerged in recent weeks that has only increased the tech-induced anxiety in Hollywood that began with the streaming revolution: OpenAI’s text-to-video generator Sora.
Karla Ortiz, a concept artist and illustrator known for her work on Marvel’s Black Panther and Dr Strange, said Hollywood was wrestling with “competing feelings”.
“There are two big blockbusters that are nominated for Oscars, and Hollywood loves that,” said Ortiz. “But there’s also budget cuts this year, and OpenAI’s release of Sora has got people freaked out.”
The freakout began when OpenAI, which is backed by Microsoft, showed off some very slick clips on February 15 that shocked cinematographers, animators and filmmakers with their quality.
Director and producer Tyler Perry was the first to sound the alarm, telling The Hollywood Reporter that he had put an $800mn expansion of an Atlanta studio on hold “indefinitely” after seeing Sora. “I had no idea until I saw recently the demonstrations of what it’s able to do,” he said. “It’s shocking to me.”
A report released in January titled “Future: Unscripted” found that entertainment groups have been among the earliest adopters of generative AI, which is likely to lead to the “elimination of many jobs entirely”.
Nearly 65 per cent of the 300 business leaders surveyed by the study said they expected generative AI to lead to the loss of jobs in the next three years, according to the report by CVL Economics, which was commissioned by the Animation Guild and other unions representing artists in the entertainment industry.
California would account for 28 per cent of the displaced creative industry jobs, followed by New York, the report said. The figures do not include gig workers and freelancers, who would also probably be significantly affected.
Brandon Jarratt, a technical director at Walt Disney animation studios and member of the animation guild’s AI task force, said there had been a range of reactions in his profession to the emergence of generative AI. “Some think it may eliminate some of the drudgery from their jobs and make it easier,” he said. “And there are other people who view it as an existential threat to the skills they spent much of their life honing.”
Right now, Sora is not ready for its close-up. It is only capable of generating about a minute’s worth of video, and it lacks a human’s understanding of physics. Falling glass is just as likely to bounce when it hits the ground as it is to break.
“There’s literally no story in these short clips, and we’re only seeing the ones that look amazing,” said Nick Lynes, co-chief executive and founder of Flawless, which makes AI tools used by filmmakers and studios. “They are wonderful examples of how the technologies are evolving but it’s going to take a long time before it’s really capable of supporting filmmakers properly.”
But Sora is learning — and how it is learning is a major concern for many in Hollywood and beyond. Entertainment companies, media groups and some individuals are trying to prevent AI companies from mining vast amounts of copyrighted material to enhance their models.
In December The New York Times became the first big US media group to sue OpenAI and Microsoft over their AI chatbots. The Times is seeking billions of dollars in damages in the lawsuit, which Microsoft has labelled “doomsday futurology” in its filing to dismiss the case.
Ortiz last year sued AI generators Stability AI, Midjourney and DeviantArt for allegedly teaching the tools from her artwork. She has similar concerns about Sora.
“When I looked at Sora, two things went through my mind: what is this trained on? Will we ever get to know?” she said. “And the other was, this is a risk to our jobs. It uses our work to power these models, and then competes against us in our own markets.”
Last year the Writers Guild of America secured groundbreaking protections against AI in its three-year contract with the Hollywood studios. Jarratt said the Animation Guild would seek similar measures as it negotiates a new deal this year, which he believed would be more effective than working through the courts.
“We are not going to be saved by copyright law,” he said. “We can try to build protections into contracts, [which] we believe we can make the most progress. Courts move at glacial speed while the technology continues to go much faster. There’s not going to be some magical ruling that just renders all of these technologies illegal.”
OpenAI said: “The training data is from content we’ve licensed and publicly available content.”
Lynes sees AI as another step in a long line of innovations that have changed how movies and TV shows are made. He says the efficiencies of AI will ultimately result in an “unprecedented” amount of new work being produced — and make “filmmaking a sound investment and sound business proposition”.
“I actually think that we’re heading into boom times in all filmmaking domains,” said Lynes. “The tools that allow for a reduction in production costs and doing things in a more efficient way are the same things and get more creative control back to filmmakers.”
Read the full article here
