OpenAI is eradicating one of many voices utilized by ChatGPT. Customers discovered that it sounded just like Scarlett Johansson, the corporate introduced on Monday, and Johansson herself launched a press release saying she employed authorized council to inquire concerning the Sky voice and get precise particulars about the way it was developed. Sky is now being paused after OpenAI used it when demoing its new GPT-4o mannequin final week.
“We consider that AI voices shouldn’t intentionally mimic a star’s distinctive voice—Sky’s voice shouldn’t be an imitation of Scarlett Johansson however belongs to a unique skilled actress utilizing her personal pure talking voice,” the corporate wrote in a weblog submit. “To guard their privateness, we can not share the names of our voice skills.”
A video of the demo made the rounds on social media final week as customers discovered the voice to be just like Johansson’s. Some mocked the voice for being overly flirtatious, whereas others likened it to a male fantasy.
The flirty voice has drawn comparisons to the 2013 movie “Her,” through which Johansson voices a sultry digital assistant. The movie’s primary character, performed by Joaquin Phoenix, finally ends up falling in love with the digital assistant.
Though the corporate hasn’t in contrast Sky’s voice to Johansson’s, OpenAI CEO Sam Altman tweeted the phrase “Her” following the corporate’s occasion.
TechCrunch confirmed that Johansson launched a press release in a while Monday saying that OpenAI approached her in September about supplying the voice, which was first reported by NPR correspondent Bobby Allyn. Within the assertion, she stated that she declined “after a lot consideration and for private causes” and was “shocked” when she heard the launched demo.
The assertion additionally says that she was “compelled to rent authorized counsel” and that the corporate “reluctantly agreed” to vary the Sky voice.
OpenAI’s demo final week aimed to showcase the chatbot’s enhanced conversational talents, however went viral after the sultry voice giggled at nearly every thing an OpenAI worker was saying. At one level, the chatbot instructed the worker: “Wow, that’s fairly the outfit you’ve acquired on.” At one other level, the chatbot stated “Cease it, you’re making me blush” after receiving a praise.
In its weblog submit, OpenAI says it desires its chatbots’ voices to sound “approachable” and to “encourage belief.” It additionally desires them to have a “heat, partaking, confidence-inspiring, charismatic voice.”
Going ahead, OpenAI says it plans to “introduce extra voices in ChatGPT to higher match the varied pursuits and preferences of customers.”
Johansson’s assertion in full:
“Final September, I obtained a proposal from Sam Altman, who wished to rent me to voice the present ChatGPT 4.0 system. He instructed me that he felt that by my voicing the system, I might bridge the hole between tech corporations and creatives and assist shoppers to really feel snug with the seismic shift regarding people and AI. He stated he felt that my voice could be comforting to individuals.
After a lot consideration and for private causes, I declined the supply. 9 months later, my buddies, household and most people all famous how a lot the most recent system named “Sky” appeared like me.
After I heard the launched demo, I used to be shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily just like mine that my closest buddies and information shops couldn’t inform the distinction. Mr. Altman even insinuated that the similarity was intentional, tweeting a single phrase “her” – a reference to the movie through which I voiced a chat system, Samantha, who kinds an intimate relationship with a human.
Two days earlier than the ChatGPT 4.0 demo was launched, Mr. Altman contacted my agent, asking me to rethink. Earlier than we might join, the system was on the market.
Because of their actions, I used to be compelled to rent authorized counsel, who wrote two letters to Mr. Altman and OpenAI, setting out what they’d completed and asking them to element the precise course of by which they created the “Sky” voice. Consequently, OpenAI reluctantly agreed to take down the “Sky” voice.
In a time once we are all grappling with deepfakes and the safety of our personal likeness, our personal work, our personal identities, I consider these are questions that deserve absolute readability. I sit up for decision within the type of transparency and the passage of acceptable laws to assist make sure that particular person rights are protected.”
We’ve reached out to OpenAI for remark and can replace this story if we hear again.
Replace: This story initially printed at 8 a.m. PT Monday and was up to date to incorporate Johansson’s assertion.