Actors in Hollywood who are now on strike worry that artificial intelligence will take their jobs, but for many stunt performers, this dystopian threat is already there.
Cost-cutting studios have used computer-generated backdrop figures for fight sequences for a long time, from “Game of Thrones” to the most recent Marvel superhero films.
Since AI is becoming more prevalent, it is now possible to construct highly complicated action sequences like vehicle chases and shootouts without the bothersome (and expensive) involvement of humans.
The long-standing Hollywood legacy of stunt work, which has appeared in everything from silent epics to Tom Cruise’s most recent “Mission Impossible,” is in danger of disappearing altogether.
“The technology is exponentially getting faster and better,” said Freddy Bouciegues, stunt coordinator for movies like “Free Guy” and “Terminator: Dark Fate.”
“It’s really a scary time right now.”
Stunt and background actors are already required by studios to participate in high-tech 3D “body scans” on location, frequently without being informed of how or when the images would be used.
These likenesses might be utilized to produce intricate, uncannily accurate “digital replicas,” which can carry out any action or speak any phrase, thanks to advancements in AI.
Bouciegues fears producers could use these virtual avatars to replace “nondescript” stunt performers — such as those playing pedestrians leaping out of the way of a car chase.
“There could be a world where they said, ‘No, we don’t want to bring these 10 guys in… we’ll just add them in later via effects and AI. Now those guys are out of the job.”
But according to director Neill Blomkamp, whose new film “Gran Turismo” hits theaters August 25, even that scenario only scratches the surface. The role AI will soon play in generating images from scratch is “hard to compute,” he told AFP.
“Gran Turismo” primarily uses stunt performers driving real cars on actual racetracks, with some computer-generated effects added on top for one particularly complex and dangerous scene.
However, according to Blomkamp, AI will get to the point where it can produce photo-realistic imagery, such as fast crashes, in as little as six or twelve months.
This is when “you take all of your CG (computer graphics) and VFX (visual effects) computers and throw them out the window, and you get rid of stunts, and you get rid of cameras, and you don’t go to the racetrack,” he told AFP.
It really is different.
