AI Voice Rights Battle: AI companies accused of stealing actors’ voices

As NPR reports,

Lawyers for Scarlett Johansson are demanding that OpenAI disclose how it developed an AI personal assistant voice that the actress says sounds uncannily similar to her own.

The lawyers are demanding details about how and why OpenAI’s AI “Sky” voice sounds like Johansson.

Johansson was the world's highest-paid actress in 2018 and 2019, known for playing Natasha Romanoff in the Marvel Cinematic Universe, among other roles. She has twice been nominated for acting Oscars.

Johansson voiced an AI character named Samantha in the 2013 movie Her. The film had only limited financial success, earning $47.4 million at the worldwide box office, but was critically acclaimed, nominated for five Oscars and winning the one for Best Original Screenplay.

“Sky” is one of five integrated voices in the AI tool called GPT-4o. Shortly before the release of GPT-4o, OpenAI CEO Sam Altman tweeted the single word "Her."

The New York Times reported that Johansson released a statement that had been approached by Altman but turned down his offer to voice the AI:

Last September, I received an offer from Sam Altman, who wanted to hire me to voice the current ChatGPT 4.0 system. He told me that he felt that by my voicing the system, I could bridge the gap between tech companies and creatives and help consumers to feel comfortable with the seismic shift concerning humans and A.I. He said he felt that my voice would be comforting to people. After much consideration and for personal reasons, I declined the offer. Nine months later, my friends, family and the general public all noted how much the newest system named ‘Sky’ sounded like me.

Johansson said that when she heard the AI demo she was “shocked, angered and in disbelief” that Altman had used a voice that sounded so “eerily similar” to hers that her closest friends and news outlets couldn’t tell the difference.

OpenAI has since turned off the “Sky” voice.

In related news, two voice actors have filed a class action lawsuit accusing AI company LOVO of stealing their voices to create millions of voiceover productions.

According to the complaint,

LOVO is a technology company with proprietary software driven by artificial intelligence (AI) that allows LOVO’s clients to create and edit voice-over narrations adapted from real actors. To be clear, the product that customers purchase from LOVO is stolen property. They are voices stolen by LOVO and marketed by LOVO under false pretenses: LOVO represents that it has the legal right to market these voices, but it does not.

The complaint also notes that “the voice-over industry is estimated to generate more than $2 billion annually in the United States, and more than $4 billion annually worldwide.”

The complaint explains that

Traditionally, actors are hired to read scripts, which may be recorded in outside studios or with the actor’s own equipment. The recordings are then edited, changes or re-recordings produced, and a final recording is laid into the show or presentation.

The actors are paid a negotiated amount for the use of their voices, and typically for the time spent recording the audio requested. The negotiated amount is a function of the actor’s name, brand value, where, and in what medium the voice-over will be used, and for how long it will be used. Payments typically include upfront fees, royalties, residuals, or some combination of these payments. And there are other potential fees required under Screen Actors Guild – American Federation of Television and Radio Artists (SAG-AFTRA) contracts.

Actors like the Plaintiffs, they claim, get paid anywhere from $150 for a short recording for a local television tag, to $2,000 for a one-time sales presentation, to $6,000 minimum for a 13-week run of a TV commercial, or more depending on the project.

Established actors can typically make much more via residuals.

According to the complaint,

Defendant LOVO, by its own admission, is attempting to disrupt and “revolutionize” this traditional model. LOVO sells a text-to-speech subscription service that allows its clients – typically companies – to generate voice-over narrations at a fraction of the cost of the traditional model. LOVO does this by allowing subscribing customers to upload a script into its AI-driven software known as “Generator” or “Genny,” and generate a professional-quality voice-over based on certain criteria. For example, LOVO customers can choose between – and designate their preference for – male or female voices, regional accents, and older or younger-sounding voices.

As of early 2023, LOVO had allegedly created over 7 million voice-overs.

The Plaintiffs were hired on the Fivrr freelance platform. Plaintiff Paul Lehrman was paid $1,200 and told by the client (who turned out to be a LOVO employee):

We are researching speech synthesis with different accents and voices. Your voiceover will be used for academic research purposes only.

Only years later did the Plaintiff learn he’d actually been doing work for LOVO.

Worse, claimed the Plaintiff,

On or about April 6, 2022, Plaintiffs learned that iNTECH, a YouTube channel (now called Military News) with more than 336,000 subscribers, had created and was promoting videos about Russian military equipment that used Plaintiff Lehrman’s AI-generated voice. To be clear: Mr. Lehrman never recorded the YouTube videos; they were generated by a then-unidentified AI software without Mr. Lehrman’s participation or approval.

Later,

on or about June 13, 2023, Mr. Lehrman heard his voice being used on a podcast episode of “Deadline Strike Talk.” Ironically, the episode was about the dangers of AI technologies.

Plaintiff Linnea Sage had a similar experience but was paid only $400.

According to the complaint,

in June of 2023, Ms. Sage discovered that LOVO had been using, manipulating, and editing her voice in promotional materials for Defendant LOVO for many years, including but not limited to, in a five-minute investor presentation at the Berkeley SkyDeck Demo Day Spring 2020 event, which was also posted to YouTube (despite previous assurance that the use of Ms. Sage’s voice would not be public). This event was used to help raise money for LOVO.

The Plaintiffs in the LOVO case are seeking at least $5 million in damages.