OpenAI has temporarily stopped the use of the ‘Sky’ voice in its ChatGPT model, following allegations from Hollywood actress Scarlett Johansson that the AI’s voice bore resemblance to hers without her consent. The controversy arose after CEO Sam Altman was accused of seeking to license Johansson’s voice, a request she declined.
In response to the allegations, Altman issued a statement stating that OpenAI had no intention of replicating Johansson’s voice. He clarified that a voice actor had already been selected before any contact was made with Johansson.
A report published in The Washington Post shed light on the process behind the creation of the AI voices. According to the report, OpenAI had issued a casting call seeking non-union actors aged between 25 and 45 for a confidential project aimed at imbuing ChatGPT with a human-like voice.
Documents and recordings cited in the report indicated that an actress had been hired months prior to Altman reaching out to Johansson. The chosen actress, reportedly not Scarlett Johansson, adopted the pseudonym ‘Sky’ to convey a soothing and refreshing tone, said the report.
Joanne Jang, product lead at OpenAI responsible for overseeing AI model behaviour, said the company meticulously selected actors enthusiastic about the project. Prospective actors were required to showcase a sample AI version of their voice to exemplify how realistic the technology could sound.
Reacting to the development, the voice artist involved in the project said, “While that was unknown and honestly kinda scary territory for me as a conventional voice over actor, it is an inevitable step toward the wave of the future.”
What really happened?
On May 19, OpenAI introduced five unique voice avatars for its ChatGPT platform, namely Breeze, Cove, Ember, Juniper, and Sky. The company stated that these voices were crafted in collaboration with professional voice actors. However, some listeners noted similarities between the Sky voice and that of Scarlett Johansson.
Also Read
According to a report by Deadline, Johansson issued a statement where she revealed Altman made him an offer to become the voice of ChatGPT 4.0.
“He told me that he felt that by 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 AI. He said he felt that my voice would be comforting to people,” Johansson said in the statement. Altman was perhaps right but the actor saw it differently and declined the offer “for personal reasons”, Johansson said.
Yet, nine months later Sky sounded very much like the Black Widow star. “When I heard the released demo, I was shocked, angered and in disbelief that Altman would pursue a voice that sounded so eerily similar to mine that my closest friends and news outlets could not tell the difference,” she added.
What made it worse for Johansson was when he tweeted ‘Her’ which was in reference to the actress’ movie. Altman, according to the actress, contacted her agent two days before the ChatGPT 4.0 demo was released but before the two parties could connect, the system was “out there”.
What did the actress do?
Johansson hired legal counsel and wrote two letters to Altman and his company OpenAI, seeking clarification on the development process of the Sky voice.
“Consequently, OpenAI reluctantly agreed to take down the ‘Sky’ voice,” she said, adding that when deepfake is such a big issue, it is pertinent to protect “our own likeness, our own work, our own identities”.
What did OpenAI say?
OpenAI categorically denied that it imitated Johansson’s voice. “We believe that AI voices should not deliberately mimic a celebrity's distinctive voice — Sky’s voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice,” it said.
Meanwhile, the voice has been taken down as of now and Johansson isn’t pursuing further legal action.