The maker of a new documentary about Anthony Bourdain says he used AI to re-create Bourdain's voice and synthesized the audio of three quotes from the TV host
‘We can have a documentary-ethics panel about it later’ — In a new documentary, Roadrunner, about the life and tragic death …
The VergeDave Gershgorn
Context & Ripple Effects
Roadrunner’s use of a synthesized Bourdain voice was not disclosed, turning a production choice into a dispute over authorship and audience transparency. The immediate follow-up coverage centered consent and disclosure as the core safeguards for synthetic media.
The case sits at an early boundary between restoration-style uses of voice technology and more expansive reuse: later coverage documented both an AI recreation of George Carlin for a comedy special and AI voice work for film localization.
First-order effects
Roadrunner’s director and producers face scrutiny over using Bourdain’s recreated voice without disclosure, while audiences must reassess whether quoted narration is archival or synthesized.
Documentary makers using synthetic performances now have a clearer reputational exposure: undisclosed fabrication can become the story rather than a seamless editing tool.
Second-order effects
Consent and disclosure become practical requirements for producers, estates, and distributors deciding whether a recognizable person’s voice can be recreated and how that use is labeled.
Voice-AI providers serving filmmakers gain a commercial use case in localization, but the Bourdain controversy ties adoption to provenance and approval practices rather than technical realism alone.
Third-order effects
Synthetic voice use is moving toward a governed-content model in which rights, consent, and disclosure determine which archives can be commercially reused.
As recreations extend from documentary narration to new performances, the distinction between preserving a voice and creating speech in its name becomes a durable media-ethics and policy fault line.
The trend: Media companies are commercializing synthetic voices while being pushed to attach clearer consent and disclosure rules to recognizable human likenesses.
When I wrote my review I was not aware that the filmmakers had used an A.I. to deepfake Bourdain's voice for portions of the narration. I feel like this tells you all you need to know about the ethics of the people behind this project. https://www.newyorker.com/... https://twitte…
The new Anthony Bourdain documentary didn't have audio of him reading emails, so they created a fake A.I. model of his voice...and didn't bother disclosing that in the film. We need a serious check on ethics in documentary filmmaking. https://twitter.com/...
Recreating stuff in documentaries usually seems like a mistake—at least if it's not obviously a recreation—and AI doesn't help matters. https://www.theverge.com/...
This blows, even though it's using his own words. It's telling that they did this for something so small. The film wouldn't fall apart without these line reads! A year ago, they would've managed. (Also that final line is such an indictment of entertainment industry panel culture)…
You can hear a line of the AI Anthony Bourdain deepfake voice in the trailer for the documentary, here it is. All credit to @hels for having the instinct to ask where the filmmaker got this audio, I'm dying to learn more about it https://www.theverge.com/... https://twitter.com/.…
when i read this I'm fascinated that Neville believes that there is a moral debate here at all. not only that but a moral debate is a quaint thing we can get round to later if we have time. Bourdain didn't say these things, didn't intend to say these things. they aren't real http…
Fans (=me included) are understandably upset over an “AI” that recreated #Bourdain's voice for the documentary. https://www.techmeme.com/... This should never be done. If it is, viewers should *always* be INFORMED that voice is reconstructed I think I'll pass on the doc
Between this, the Whitney/Tupac hologram, & NFTs, even non-"famous" people are now gonna have to come up with elaborate & ironclad image & likeness clauses in their wills all because YT people don't know how to leave you tf alone—even when you're dead 🤨 https://newyorker.com/...
Excited to someday feed a bunch of someone's past statements into GPT-9 and have it generate authentic-sounding quotes on subjects that had never crossed their mind For example, what would Ronald Reagan have thought about Weezer Until now, the world has never known https://twitte…
I finished @hels piece on the Bourdain doc and you should read it and make your decision, but this made me decide I'm not watching it. I also already got to read the oral history and it was already intense enough and I'd suggest waiting for that. https://www.newyorker.com/...
Faking Bourdain's voice? Is it appropriate to deepfake a dead person's voice? Under what circumstances and with what disclosure? We've been working on ethics principles around deepfakes @witnessorg. Consent, deceptiveness, disclosure, audience expectations in genre all key.. http…
There's a fascinating ethical dilemma about post-death performances, which celebrity culture has been at the forefront of for a while. @DeathandCulture is the expert on this https://twitter.com/...
for what it's worth, the unease about this pretty directly echoes the contemporary unease about the invention of recorded sound “allowing the dead to speak” https://twitter.com/...
Can never know for sure, but it certainly does *feel* like the kind of thing the subject of the documentary would have thought was deceptive or disingenuous https://twitter.com/...
That Anthony Bourdain doc used AI to deepfake his voice and make Bourdain posthumously say things he didn't say. That's ethically fucked up on so many levels and is really gross. https://www.inputmag.com/...