Actor James Earl Jones will no longer voice Star Wars villain Darth Vader in forthcoming film or television projects.
Jones, 91, has lent his voice to Vader for 45 years, first voicing the Sith Lord in Star Wars back in 1977.
But with his voice deteriorating with age, Jones will no longer appear in the Star Wars universe.
Rather, Jones has signed a deal that allows Lucasfilm (owned by Disney) to recreate his voice using artificial intelligence technology in future projects.
Disney has enlisted Ukrainian AI program Respeecher for the task, with its software drawing on years of archived audio files to recreate Jones’ iconic delivery.
Sound editor Matthew Wood said Jones last voiced Vader in 2019’s film The Rise of Skywalker.
Even then, he said Jones was ready to step back, prompting Wood to start thinking about how to “keep Vader alive and vital”.
“He had mentioned he was looking into winding down this particular character. So how do we move forward?” Wood told Vanity Fair.
It was decided to use Respeecher’s AI tech in the Disney+ series Obi-Wan Kenobi (for which Jones was credited) – with the producers describing it as a “moment of celebration”.
Now Respeecher’s responsibility for Vader’s voice has become permanent.
Jones and Disney’s arrangement may seem unconventional, but it is part of a wider trend.
AI is being used for everything from audiobooks to the dubbing of foreign films.
Even video game dialogue and music vocals are increasingly generated that way – making voice actors increasingly redundant.
Voice actors are all too aware of this uncomfortable reality.
A survey taken earlier this year by Equity, the British union for the performing arts, found that 65 per cent of its members thought AI posed a threat to their jobs.
“I am interested in what AI can do, but its capacity for imitating real people is alarming,” Red Dead Redemption 2 voice actor Roger Clark said earlier this year.
Gizmodo writer Justin Carter is also concerned.
“Similar to discussions about AI art, there’s something ethically dicey about this, and that feeling will likely not go away as Disney uses it in other shows or films,” Carter wrote this week.
How do they do it?
The team behind Respeecher uses a combination of archival recordings and an AI algorithm to create new dialogue.
Along with creating a “perfect match” of someone’s voice from existing recordings, Respeecher can reproduce speech patterns and delivery.
Its technology can also be used to transform a person’s delivery from an imported performance.
For example, voice recordings can be distorted to sound like they were spoken by someone of a different gender or different age.
Respeecher said it can generate dialogue from a person speaking in another language.
It is not the only company harnessing the power and possibilities of artificial intelligence in the world of entertainment.
Sonantic is a British AI company that was recently snapped up by streaming giant Spotify. It, too, claims to create life-like, AI-generated voices.
Resemble AI is a similar service, readily accessible on a web browser, which generates human-like voices “in seconds”.
Resemble AI said it can change the emotions in a voice recording, to make someone sound happy or sad or angry.
Cutting costs
Many film studios have turned to companies like Respeecher to cut costs and streamline production – to the dismay of voice actors.
In the past, studios typically brought voice actors into booths to record ‘filler’ for background audio; AI programs have removed this step.
These days, instead of calling an actor back into the studio if lines need to be redone, AI can generate them from existing audio files.
Many studios already use AI technology for such tasks.
To dub foreign movies, for example, AI is used to generate dialogue in different languages with the press of a button.
The ‘dark side’
Although James Earl Jones has given Disney permission to regenerate his voice for future projects, others haven’t been as fortunate, with the issue of consent a hot topic.
Fans of Anthony Bourdain shared their anger in 2021 when it was found that the late celebrity chef’s voice was artificially generated in a documentary about his life.
Roadrunner: A Film About Anthony Bourdain, included a 45-second audio clip, in which Bourdain appeared to read out an email sent to a friend.
However, director Morgan Neville did not disclose his use of AI to viewers. He eventually admitted that he had done so – without permission from Bourdain’s family or estate – to the concerns of fans and critics.
Legal action
The use of AI-generated voices has also led to legal action.
Voice actor Bev Standing sued ByteDance, the company behind social media platform TikTok, in 2021 when she discovered her voice was being used for the platform’s ‘text-to-speech’ feature.
The feature, which allowed users to type text into the platform and then generates audio artificially spoken by Standing, became available in 2020.
Standing said she made recordings for a text-to-speech feature for the Institute of Acoustics, part of the Chinese Academy of Sciences, years before.
She was unsure how they ended up in ByteDance’s possession. The institute has denied giving or selling them to ByteDance.
Her voice soon became an inescapable presence on the platform, heard in TikToks far and wide.
TikTok agreed to settle Standing’s lawsuit weeks after it was filed, subsequently changing the voice used in its function.
