Fable, the AI startup backed by major investors, is pushing ahead with its controversial plan to recreate 43 minutes of lost footage from Orson Welles' 1942 masterpiece "The Magnificent Ambersons" using generative AI. The project, which drew skepticism when announced last fall, now has the cautious blessing of Welles' estate and backing from a prominent biographer—but critics still argue that recreating the destroyed scenes fundamentally misunderstands what makes art meaningful. As TechCrunch reporter Anthony Ha puts it, the initiative raises fundamental questions about AI's role in preserving versus distorting creative legacy.
When Fable founder Edward Saatchi announced plans last fall to resurrect lost scenes from Orson Welles' "The Magnificent Ambersons," the reaction from film purists was swift and brutal. But a new profile in The New Yorker reveals a more complex story—one driven by genuine cinephilia rather than Silicon Valley hubris, even as it raises uncomfortable questions about AI's expanding reach into creative preservation.
Saatchi, whose father co-founded the legendary advertising firm Saatchi & Saatchi, grew up in a "movie mad" household with a private screening room. He first watched "Ambersons" at age twelve, and the film's tragic history has haunted him ever since. After a disastrous preview screening in 1942, RKO Pictures slashed 43 minutes from Welles' cut, tacked on an unconvincing happy ending, and eventually destroyed the excised footage to free up vault space. Welles himself called it "a much better picture" than "Citizen Kane."
"To me, this is the holy grail of lost cinema," Saatchi told The New Yorker. "It just seemed intuitively that there would be some way to undo what had happened."
That intuition led Fable to partner with filmmaker Brian Rose, who'd already spent years attempting a similar restoration using animated scenes based on Welles' script, photographs, and production notes. Rose's earlier attempt didn't exactly wow audiences—he admitted friends and family "were scratching their heads" after viewing it. Fable's approach is more ambitious: filming live-action scenes, then overlaying them with AI-generated recreations of the original actors and their voices.
But the technical challenges are significant, even by AI standards. The company has grappled with glitches like a two-headed version of actor Joseph Cotten. There's also what Saatchi calls a "happiness problem"—the AI tends to make the film's female characters look inappropriately cheerful, undermining the story's melancholic tone. And recreating the complex cinematography that made the original so visually stunning remains a daunting task.
Saatchi conceded he made "a total mistake" by not consulting Welles' estate before going public. Since then, he's been working to repair relationships with both the estate and Warner Bros., which owns the film rights. Welles' daughter Beatrice has softened her stance somewhat, telling The New Yorker she now believes the team is approaching "this project with enormous respect toward my father and this beautiful movie," though she remains "skeptical."
The project has also won over Simon Callow, the actor and biographer currently writing the fourth volume of his definitive Welles biography. Callow, a family friend of the Saatchis, called it "a great idea" and agreed to serve as an advisor.
But not everyone's convinced. Melissa Galt, daughter of "Ambersons" actress Anne Baxter, said her mother would have opposed the project entirely. "It's not the truth," Galt explained. "It's a creation of someone else's truth. But it's not the original, and she was a purist." Her mother's position, she said, was simple: "Once the movie was done, it was done."
That philosophy cuts to the heart of what makes the Fable project so contentious. Writer Aaron Bady recently compared AI to the vampires in the film "Sinners," arguing that both fundamentally misunderstand art because they lack knowledge of mortality and limitation. "There is no work of art without an ending, without the point at which the work ends," Bady wrote. "Without death, without loss, and without the space between my body and yours, separating my memories from yours, we cannot make art or desire or feeling."
By that measure, Saatchi's insistence that there "must be some way to undo what had happened" reflects an almost childlike refusal to accept permanent loss—not unlike a studio executive deciding "The Magnificent Ambersons" needed a happy ending in the first place.
The New Yorker profile notably doesn't include any footage of Fable's AI-hybrid results, only clips from Rose's earlier animations and still images of AI-generated actors. Whether the project will ever see public release remains uncertain, pending negotiations with the estate and Warner Bros.
What's clear is that Fable's "Ambersons" experiment represents a new frontier in how AI intersects with cultural preservation. Similar debates are erupting across the industry, from startups claiming they can use AI to make grief obsolete by recreating deceased loved ones, to questions about AI's role in restoring versus reimagining lost artworks.
For Saatchi, the motivation appears genuine—a film lover's dream of glimpsing Welles' original vision. But as Galt's comments suggest, the result will likely be just that: a dream, a novelty, rather than the lost masterpiece itself.
The 'Magnificent Ambersons' project crystallizes AI's thorniest challenge in the creative space—the difference between preservation and resurrection. While Saatchi's passion for Welles is genuine and he's secured cautious support from key stakeholders, the fundamental question remains: Can AI meaningfully restore what was lost, or does it only create an elaborate simulacrum? As the technology advances and similar projects proliferate, the industry will need to grapple with whether some losses should remain permanent—not as failures, but as intrinsic to what makes art human in the first place. For now, Fable presses forward with its experiment, and cinema lovers will be watching closely to see whether the results justify the controversy.
