Steven Soderbergh’s latest documentary, John Lennon: The Last Interview, premiered at the 79th Cannes Film Festival, sparking intense debate—not over John Lennon’s final words, but over the filmmaker’s decision to use artificial intelligence for roughly 10% of the visuals. The AI-generated sequences, created using Meta’s proprietary software, drew sharp criticism from festival attendees and film critics, but Soderbergh remains unapologetic. In fact, he argues that the firestorm of condemnation is exactly the point: the real problem is that almost no one else in Hollywood is admitting they use AI at all.
The documentary is built around a never-before-released two-hour-and-45-minute radio interview that Lennon and Yoko Ono gave to San Francisco’s KFRC radio crew on December 8, 1980, just hours before Lennon was shot and killed outside the Dakota Apartments. Soderbergh and co-editor Nancy Main condensed the audio to 97 minutes, creating what has been described as a hyperkinetic photo album stitched together from more than 1,000 archival photographs and video clips. But for moments when the conversation turns philosophical—discussing love, creativity, or the nature of fame—Soderbergh turned to Meta’s generative AI to produce abstract imagery where no archival footage existed.
The AI sequences: abstract, surreal, and controversial
The AI-generated sections are deliberately non-representational: circles of light, a black rose morphing into a choreographic pattern, paint colors mixing in split screen as lovers caress. There are no deepfakes of Lennon or fabricated historical images. Soderbergh told the Associated Press that he was fully prepared for the backlash. “I knew what was coming,” he said. “You don’t say yes to Meta offering you these tools and offering to finish the film and not know you’re going to come in for some heat. That was part of the deal.” The response he received was exactly what he expected. Critics at Cannes described the AI sequences as the weakest part of an otherwise immersive and emotionally powerful film.
Soderbergh’s defense rests on two main arguments: necessity and transparency. He insists that the surreal visuals could not have been produced through conventional visual effects within the film’s modest budget. “It has to be necessary. Is it the only way to accomplish what I want to see? Is it truly the best way to do it?” he asked. He admitted that he struggled to articulate what he wanted to see to his collaborators. “I wasn’t very articulate to the people I was working with,” he said. “It was hard to describe the things I wanted to see. The good part about this technology was at least the ability to have something in front of me quickly that I could respond to.” The ability to iterate rapidly, he explained, allowed him to refine imagery that would otherwise have remained locked in his head.
But the broader argument is about honesty. Soderbergh points out that while he voluntarily disclosed the partnership with Meta, countless other filmmakers and advertisers are using AI without any public acknowledgment. “In the world outside of the creative context, we’re not aware of the extent that this is being used and used to manipulate us,” he said. “We don’t know because they’re not telling. We find out after, by accident, by some whistle blower. I’m like my own whistle blower.” His stance is deliberately provocative: by flagging his own AI use, he hopes to force a conversation about the industry’s structural dishonesty.
Industry context: AI adoption outpaces disclosure
Data from Canva’s State of Marketing and AI Report, released the same week as the Cannes premiere, underscores Soderbergh’s point. The report found that 97% of marketing leaders now use AI daily, yet 78% of consumers still prefer human-made creative work, and 87% believe the best advertising requires a human touch. Mentions of “AI slop” have increased ninefold, reflecting a growing consumer distaste for AI-generated content—even as companies deploy it behind the scenes. The gap between how widely AI is used and how willingly creators admit to it is the very dishonesty Soderbergh is calling out.
The film industry has been cautiously integrating AI for several years. Flawless AI’s DeepEditor, which digitally adjusts actors’ lip movements to match dubbed audio, has been used in mainstream productions since 2022, with performer consent managed through the Artistic Rights Treasury platform. The 2023 SAG-AFTRA strike established that any meaningful digital alteration of a performance requires explicit actor consent. Soderbergh’s case is different: he is not altering an existing performance but generating entirely new visual content to accompany audio that has no corresponding video. The ethical territory is less charted, and Soderbergh acknowledges that he is still exploring his own boundaries. “I don’t know where my line is yet. I’m waiting to see,” he said. “Each creative person is going to have their own prism and be affected by it in different ways. Our inherent desire to have a simple template for how this is to be approached is part of the problem. I don’t think that’s possible.”
Soderbergh’s views on AI’s impact on filmmaking jobs are surprisingly measured compared to many industry voices. Rather than seeing AI as a job-killer, he suggests that as the technology makes technical perfection accessible to anyone, human imperfection will become the scarce and therefore valuable commodity. “As it becomes possible for anybody to create something that meets a certain standard of technical perfection,” he said, “then imperfection becomes more valuable and more interesting.” This formulation inverts the usual anxiety: instead of AI raising the floor and eliminating human work, Soderbergh believes it will highlight the irreplaceable quality of human messiness and unpredictability.
The documentary itself: a powerful portrait despite the controversy
Critics are generally in agreement that the documentary is a powerful work, regardless of the AI controversy. The Wrap called it a film that “does as much to demystify Lennon and Ono as ‘Get Back’ did to the Beatles.” Variety acknowledged the AI sequences as the weakest part but praised the overall immersive experience. The captured conversation, edited down from 165 minutes, shows Lennon at 40 in a state of unusual clarity and openness. He speaks about love, parenthood, creativity, and his desire to dismantle the “male rock star myth” at a time when no other major rock star was interested in doing so. Soderbergh said, “What I hope young people who see it get out of it is: This guy told the truth about everything from the jump, right up through the last day of his life. He was very opinionated but also very thoughtful and all in the aid of: Can we do this better? Can we do a better version of human beings on this planet?”
Soderbergh’s career has been marked by a restless experimentation with form and technology. From his breakout film sex, lies, and videotape (1989) to the sleek heist trilogy Ocean’s Eleven, from the experimental Full Frontal to the iPhone-shot Unsane, he has consistently pushed boundaries. His willingness to wade into the AI debate is consistent with his role as a filmmaker who thrives on provocation and inquiry. This documentary, financed in part by Meta—which provided both the AI tools and the funding to complete the project—may shape the conversation around AI in documentary filmmaking for years to come.
The film does not yet have a distributor. Whether audiences beyond Cannes will have the chance to judge the AI sequences for themselves, or whether the controversy will overshadow the conversation it was built to preserve, remains an open question. What is clear is that Soderbergh has succeeded in one goal: making the industry talk openly about how and when AI should be disclosed. The debate he started may outlast the documentary itself.