3 min Analytics

Actress Scarlett Johansson claims OpenAI chatbot voice is ‘eerily similar’ to her own

Actress Scarlett Johansson claims OpenAI chatbot voice is ‘eerily similar’ to her own

Actress Scarlett Johansson accuses OpenAI of making one of the voices of the Voice Chat functionality in ChatGPT ‘eerily similar’ to her own voice. She made this statement after OpenAI removed this voice type, named ‘Sky’, as an option.

According to the Hollywood actress, OpenAI CEO Sam Altman approached her last September, asking if she would ‘lend’ her voice for one of the voice options users can choose from during Voice Chat. She declined that offer but says she recently discovered how similar the voice of the Sky chatbot was to her own.

She comments, “When I heard the released demo, I was shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine that my closest friends and news outlets could not tell the difference.” In a response, Altman states that the voice used is not Johansson’s, but that of another voice actress. The resemblance would be unintentional, he told Reuters.

Four voices left for AI chatbot

Nevertheless, OpenAI removed the voice. Of the five original voices, Breeze, Cove, Juniper, Ember and Sky, available in the latest version of ChatGPT, only the first four now remain. Those who choose Sky (while they still can) will hear Juniper. OpenAI’s Voice Chat FAQ states that the use of the Sky voice variant has been paused.

Incidentally, according to the actress, Altman allegedly insinuated in a message on X that the voice used was indeed hers, or at least purposefully similar to hers. On May 13, he sent a message with only the word ‘her,’ which allegedly referenced the 2013 film Her (written without capital letters, just like on the movie poster). In it, actor Joaquin Phoenix plays a man who enters into a relationship with Samantha, a female-sounding AI. Scarlett Johansson played this AI.

‘Alignment team’ disbanded

Last week, OpenAI disbanded the ‘alignment team’ that oversaw security issues regarding AI. In short, this team checked whether AI development was still aligned with human needs and priorities. The remainder of this team will now be integrated into other research teams, thus ceasing to exist as an independent entity, Bloomberg reports. This news followed shortly after the departure of the two team leaders, Ilya Sutskever and Jan Leike. Both are veterans of OpenAI; in fact, Sutskever is one of the company’s co-founders.

The two created the team last year, partly because of concerns about the speed of AI development. The pair regularly clashed with CEO San Altman, who has somewhat fewer inhibitions in this area. Leike confirmed that the alignment team regularly had to ‘sail against the wind’. Earlier, other members of this team said goodbye to OpenAI as well.

Also read: ChatGPT now talks in real-time, new GPT-4o model available for free