New Anthony Bourdain Documentary Includes Controversial AI-Generated Voiceover
The new Anthony Bourdain documentary Roadrunner is intercut by director Morgan Neville with audio of the famous chef narrating parts of his life. Some of those clips are not actual recordings of Bourdain, but rather written quotes transformed into synthetic speech. It might have gone unnoticed if Neville hadn’t mentioned it in an interview with the New Yorker, but the deepfake voice is now at the center of a heated debate over whether he had approval from Bourdain’s widow and if the computer-generated audio makes the documentary fraudulent.
Neville said there are three moments where the synthetic voice is used, but didn’t give specifics of when in the movie they are heard. He also didn’t share what specific products he used to devise it, but there are an ever-growing number of companies capable of taking what Neville said was a dozen hours of Bourdain speaking and analyzing it to produce a model that can read out Bourdain’s writing while sounding very like him. The debate over whether Bourdain would have approved the artificial voice had he been asked before his suicide is complicated by an argument between Neville and Bourdain’s widow Ottavia Bourdain. Neville claimed that he checked with her and Bourdain’s literary executor and that they were fine with it and that Bourdain would have been fine with it as well. Ottavia Bourdain shared her strenuous disagreement with that story on Twitter.
I certainly was NOT the one who said Tony would have been cool with that. https://t.co/CypDvc1sBP
— Ottavia (@OttaviaBourdain) July 16, 2021
This specific case has several unique factors, but the accelerating improvement and spread of synthetic voice technology mean questions of rights and permissions are being worked out in several ways depending on the company. Fun projects like this fan-made trailer for Skryim built with synthetic voices become much more serious when money is involved. Big-name voice talents have their own deals for how the voices generated get used and how much money they get for it, while platforms like Lovo and Veritone are soliciting people to record their voices in their own time, with payment for if and when a brand or company uses their voice. Things get even more complicated when you throw in the hyperrealistic virtual humans made with Unreal MetaHuman Creator combined with the synthetic voices of Replica Studios. What’s happening with Neville could degenerate into a lawsuit comparable to the one filed by the woman who claims her voice recordings were used without her permission to generate the artificial voice on TikTok. The social video platform switched out the voice not long after, but the legal and financial battles are still ongoing.