Steven Soderbergh embraces AI while releasing a film about art forgery

One of too many fingers curl on a monkey's paw, as an upcoming Soderbergh historical epic promises to be full of AI.

Steven Soderbergh embraces AI while releasing a film about art forgery

In the middle of labor negotiations over Hollywood’s use of AI, director Steven Soderbergh announced he was going all-in on the generative technology. In a recent interview with Filmmaker Magazine, Soderbergh shared that he was experimenting with AI on his next film. “I’ve been working with AI lately on the John Lennon and Yoko Ono documentary that we’re almost done with,” he told Amy Taubin. “AI has been helpful in creating thematically surreal images that occupy a dream space rather than a literal space.” He explained that while the majority of the movie would be composed of archival footage and stills, “…10 minutes, spread out over the 90-minute film, are these little pockets of images we created whenever they start talking philosophically.” But Soderbergh’s not just dabbling in AI, he’s embracing it for his next project. 

In the same interview, the filmmaker said he was planning on using “a lot of AI” for his upcoming movie with Wagner Moura set during the Spanish-American War. A career-long innovator since his splashy 1989 debut Sex, Lies, And Videotape, this approach to production is perhaps not a surprise—he’s interested in making more movies in quicker succession than he’s already demonstrated through his prolific output. And sure, there are very few ways to create a large-scale historical epic as an independent filmmaker without a deep-pocketed studio backer, but should audiences have to accept Soderbergh’s AI slop as the consolation prize? 

While some filmmakers have spoken out against AI (Guillermo del Toro would “rather die” than use it), others like Soderbergh are experimenting with these tools in ways both hyper-detailed and all-encompassing. Brady Corbet went on the defensive during the Oscar campaign for The Brutalist when questions arose around the use of AI to refine its stars’ Hungarian accents. Recently, there was an immediate backlash to Darren Aronofsky’s AI-generated series On This Day…1776, where commenters pointed out typos in historical documents and other odd behaviors in the trailer. Other filmmakers have attempted to use AI with more nuance, as Raoul Peck did with AI-generated images to illustrate examples of misinformation and online racism in his documentary Orwell: 2+2=5, or Radu Jude did with his satirical Dracula. While Hollywood wrestles with the back-and-forth around AI use, internationally, a number of Asian production companies have fully incorporated AI into their production slate in an effort to cut costs and address declining ticket sales.

While many of these are predominantly business and labor decisions, film is still an aesthetic medium, and in its current form, a lot of AI-generated images look like hot garbage on a summer’s day without any additional human input. It’s an affront to the senses and odious to look at for those of us not falling for weirdly manipulative animal videos or bizarre humanized fruit serials. This deluge of distasteful imagery coincides with a reviving interest in analog and physical objects; it’s no surprise when moviegoers are disappointed to learn their new favorite film dabbled in AI, as was the case with Late Night With The Devil. In that case, the indie horror used three AI images as interstitials, which were later altered, but by that time, its word-of-mouth was already tarnished. 

Visual quality concerns aside, there have been precious few conversations about the ecological impact of sustained genAI use, especially since Hollywood––and much of the western half of the U.S.––are in water-sensitive areas, where data center usage may potentially add a strain on drinking water reserves. Then there’s genAI’s original sin: training on copyrighted material to regurgitate it back, or using the likeness of people who did not consent to being used like a digital puppet.

This all comes back to Soderbergh’s statements. His latest film, The Christophers, doesn’t directly address the AI elephant in the room, but it does question the concept of authenticity and whether or not an artist can replicate another artist’s style and pass it off as an original. Characters describe painted works being finished by assistants and forgeries slipping under the radar—isn’t it all the same if the buyer believes they’re getting the real deal? But if a forged painting is fake, what makes a fake film? Does too much AI dilute the original author’s artistry? The story, written by No Sudden Move’s Ed Solomon, ends by embracing a work of murky authorship, authenticity be damned. 

There’s no stopping professional tinkerers from playing with these new toys—even if the mushy resolution and flavorless gimmickry of generative AI images still look like leftovers that stayed in the microwave too long. Even Soderbergh admitted “you need a Ph.D. in literature to tell it what to do,” because the nascent technology still requires copious human input to refine its lumpy first drafts. According to the Academy, they don’t care if AI is used as a tool so long as “human creativity” stays at the heart of the picture, but what else are filmmakers sacrificing to rush an idea through production if not the human labor it takes to make movies? Since the viewers are where all the pushback is coming from, the next step would be to use AI in such a way that even a discerning audience can’t tell—a forgery slipped under the radar.

 
Join the discussion...
Keep scrolling for more great stories.