I first learned that “Roadrunner,” Morgan Neville’s documentary about the life and death of Anthony Bourdain, contains three sentences spoken by Bourdain that he never actually spoke out loud in the same way that you learn about a lot of things these days: by seeing an eruption of outrage about it on Twitter. The eruption immediately sent me to the New Yorker article in which Neville, the award-winning director of “Won’t You Be My Neighbor?” and “20 Feet From Stardom,” first explained how he used AI technology to feed 10 hours of Bourdain voice recordings into a computer, which then simulated Bourdain’s reading of those sentences — every one of which he had, in fact, written.
The words weren’t faked; the sound of him speaking them was. This was characterized, on social media, as an ethical lapse, and my first reaction is to say that I don’t necessarily disagree. My second reaction is to say that as ethical lapses go, this isn’t exactly a crime against humanity. The reason I think it’s a vital discussion to have, but one that (to me) provoked a bit of pearl-clutching, is that if the issue on the table is how documentaries represent and manipulate and distort reality, there has been too much inauthentic water under the bridge already. When it comes to swapping in fake reality, documentaries have been sliding down a slippery slope for years.
Neville, in the New Yorker article, defended his choice as a novel way to make Bourdain’s words come alive, adding, “We can have a documentary-ethics panel about it later.” Maybe we should have one. But maybe a reason that Neville, in that quote, adopted what will strike some as a cavalier attitude is that no one understands more than documentary filmmakers what an impure form the documentary can be. Investigations into the gray areas of documentary integrity have often gone back to the movie that planted the genre on the map, Robert Flaherty’s 1922 silent film “Nanook of the North,” which presented itself as a “documentary” portrait of Inuit life, even though most of it was staged. It used actual Inuk people and settings, but Flaherty directed the action and, in many cases, concocted it. It’s about as far from a pure-form documentary as you can get.
To be fair, “Nanook of the North” didn’t set the standard that documentaries are now measured against. The form became more authentic over time, and by the 1960s, the age of the cinéma vérité revolution, filmmakers like Albert and David Maysles, D.A. Pennebaker, Richard Leacock, Jean Rouch and Fred Wiseman had forged a stunning new art form in which the technology of portable sound-sync cameras allowed us to eavesdrop on life as it was happening. A story that generally remained untold was one about the subtle off-camera relationship between the directors and their subjects, and how that impacted the reality we were seeing. (There could be a documentary-ethics panel on that one.) Nevertheless, the vérité revolution was genuine, and it established certain codes. So did the classic-form archival doc that began to come into its own around the time that Ken Burns was making “Brooklyn Bridge” (1981). For those of us who revere documentaries, this was the start of a grand era.
But by the late ’80s, the documentary was starting to be messed with. Michael Moore, in his proletarian baseball cap, placed himself at the center of his films, turning them into a form of political performance art, and in “Roger & Me” (1989) he manipulated the chronology of events. The biggest change, though, arrived with Errol Morris’s “The Thin Blue Line” (1988), a documentary-as-murder-investigation that more or less introduced the idea of plopping staged reenactments into the middle of an otherwise nonfiction film. Morris did it for what appeared to be a grave moral purpose (exonerating an innocent man), and “The Thin Blue Line” was such an ingeniously made true-crime-story-as-documentary-noir that it seemed, at the time, to create a new art form. But what was really happening is that Morris, in taking the liberties he did, had let a genie out of the bottle.
The staged scenes in “The Thin Blue Line” didn’t bother me at the time, and they don’t now. That movie really is a one-of-a-kind landmark. What did bother me is when I began to notice how routinely documentaries were folding in staged reenactments. It first struck me around the time of “Man on Wire” (2008), James Marsh’s otherwise enthralling film about Philippe Petit’s staggering, don’t-look-down walk on a high wire strung between the two towers of the World Trade Center in 1974. Amazingly, there wasn’t any filmed footage of the event, so the documentary, at its climax, had to rely on still photographs of Petit standing on that wire.
But to tell the more prosaic story of how he and his helpers hid out in the World Trade Center, the film featured staged scenes of actors sneaking into the buildings, as if out of some Hollywood thriller. I could feel the scenes taking me out of the movie, even as their purpose was, ostensibly, to heighten our involvement. This kind of thing has now become routine, employed in documentaries as diverse as Kevin Macdonald’s harrowing mountain-climbing epic “Touching the Void” (2008), Yance Ford’s murder-investigation autobiography “Strong Island” (2017), Alex Gibney’s “Client 9: The Rise and Fall of Eliot Spitzer” (2010) and Andrew Jarecki’s influential true-crime serial “The Jinx” (2015). I wouldn’t call the staged scenes in any of these documentaries “unethical,” but I would say that they’re a way of synthesizing realities it would have been better to let us imagine in our heads. To me, staged reenactments, even though we’re all used to them, are a more dispiriting phenomenon than the faking of Anthony Bourdain’s voice.
You might say that they’re two separate things, and that reenactments don’t pretend to be genuine. But I see it differently. A reenactment and a voice fake actually do different versions of the same thing: Both cement a reality in your mind — the image of something or the sound of something — that didn’t happen, at least not in the way it’s presented. And I’d argue that the deepfake Bourdain voice probably gets closer to reality than most reenactments do. Many of those who objected to what Morgan Neville did with Bourdain’s voice argue that if the same technique had been labeled, it would have been okay. I tend to agree. But where should the labeling happen: During the movie or during the closing credits? I’ve seen good documentaries in which an actor will read a subject’s words, sometimes simulating their tone of voice (as in the superb 2006 film “Jack Smith and the Destruction of Atlantis”). It’s not much of a leap from that to what “Roadrunner” does.
I think what bothers people, justifiably, about the Bourdain film isn’t so much the “ethical lapse” as the sudden dramatic implication of how scarily reality can now be manipulated. We’re only at the dawn of the age of the deepfake. People can now be made to look like they’re doing things, or saying things, that they never did or said. The manipulation of Bourdain’s voice in “Roadrunner” seems to open a Pandora’s Box. What happens when unethical filmmakers employ such techniques? But let’s not pretend that we’ve been purists about it. Documentaries have been inching away from unalloyed reality for a long time. And it’s we in the audience who enable it. We’re the ones who like our reality sweetened, heightened, finessed until it looks just like a movie.