'Sky' Gets Called to the Sky

OpenAI pulls their 'Samantha'-like voice – under threat from Samantha herself...
OpenAI to Pull Johansson Soundalike Sky’s Voice From ChatGPT
OpenAI is working to pause the use of the Sky voice from an audible version of ChatGPT after users said that it sounded too much like actress Scarlett Johansson.

To paraphrase Bane from The Dark Knight Rises, "no one cared who I was until I put on the voice-first UI".

Thus is the story of 'Sky', one of the voices from OpenAI. While "she" has been around since the service launched its first iteration of spoken interaction within ChatGPT, it seems like everyone only really noticed her last week, when most of OpenAI's GPT-4o demos showcasing their (insanely impressive) new vocal computing capabilities used 'Sky' and yes, made some references to the movie Her. The voice assistant in Her is named 'Samantha' and is voiced by Scarlett Johansson. Though the voice of 'Sky' was apparently created with the help of another (unnamed) voice actress, it was not Scarlett Johansson.

Still, the fact that everyone keeps bringing up Johansson in this context, from The New York Times to Saturday Night Live,1 meant that OpenAI needed to take action. It seems highly unlikely that Johansson could sue – well, to be clear, she could, but it feels very unlikely she could win such a case – over this, nor was there any indication she was going to. But it's also not a great look for OpenAI to be using a voice that so clearly evokes a very famous voice in a very famous context in an age when everyone is freaking out about AI taking jobs – especially in Hollywood.

🤖
As it turns out, Scarlett Johansson was very aware and very unhappy about the use of a voice similar to her own, more below...

Still, it's too bad. 'Sky' was a nice voice! I've listened to many different synthetic voices over the years and 'Sky' seemed right up at the top.2 And now, ironically, OpenAI is sticking even closer to the script of Her by calling 'Sky' away to the great beyond just as happened to 'Samatha' in the film. I still argue that the end of that movie is not some great dunk on AI, but instead a poignant apology note between filmmakers, but OpenAI is making it harder to argue now.


Update May 21, 2024:

Scarlett Johansson says she was ‘shocked, angered’ when she heard OpenAI’s ChatGPT voice that sounded like her
The “Her” actor released a statement following OpenAI pulling its “Sky” voice from ChatGPT.

How's this for a plot twist?

"Last September, I received an offer from Sam Altman, who wanted to hire me to voice the current ChatGPT 4.0 system," Johansson wrote in a statement, which a representative shared with NBC News. "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 AI. He said he felt that my voice would be comforting to people."

"After much consideration and for personal reasons, I declined the offer," she continued. "Nine months later, my friends, family and the general public all noted how much the newest system named 'Sky' sounded like me."

As bad as the look above may have been, now this is a really bad look. But it's made about a million times worse by the fact that Altman himself referenced Her in relation to the GPT-4o system.

"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," Johansson wrote in the statement. Altman’s announcement of the new product was posted on X on the same day of the product demonstration and is still live. Her statement continued, "Mr. Altman even insinuated that the similarity was intentional, tweeting a single word 'her' - a reference to the film in which I voiced a chat system, Samantha, who forms an intimate relationship with a human."

OpenAI clearly – clearly – wanted and tried to get Johansson for the "role" and when she declined, it sure looks – and sounds – like they figured out another way to get a similar, if not exactly the same, end result.

"Two days before the ChatGPT 4.0 demo was released, Mr. Altman contacted my agent, asking me to reconsider. Before we could connect, the system was out there," Johansson wrote. "As a result of their actions, I was forced to hire legal counsel, who wrote two letters to Mr. Altman and OpenAI, setting out what they had done and asking them to detail the exact process by which they created the 'Sky' voice. Consequently, OpenAI reluctantly agreed to take down the 'Sky' voice."

Yeah, two days before is really, really, really, really not a good look here. And again, I'm not sure what legal standing Johansson has if OpenAI truly did just hire another voice who sounds similar to her – I'm sure we'll get about ten thousands lawyers weighing in on this soon enough! – but clearly OpenAI and Altman have opened themselves up to at least the threat of legal issues with their actions.

"In a time when we are all grappling with deepfakes and the protection of our own likeness, our own work, our own identities, I believe these are questions that deserve absolute clarity. I look forward to resolution in the form of transparency and the passage of appropriate legislation to help ensure that individual rights are protected," she wrote.

Legality aside, this is a PR nightmare for the pre-eminent AI company; to be under threat of lawsuit from one of the biggest actors in the world. Some may recall when Johansson got Disney to bend to her will around streaming rights during COVID. This is potentially a far bigger situation, with all of Hollywood and the larger creative community getting ready to assemble around her like the Avengers.


Update May 22, 2024: A few more thoughts/questions on the matter...

Who Is She in ‘Her’?
The curious case of Scarlett Johansson’s voice…

Update May 23, 2024: And some further clarity thanks to a few OpenAI documents and 'Sky' herself speaking!

‘Sky’ Speaks!
Documents and timelines make OpenAI’s case -- for now

1 Weekend Update anchor, Colin Jost, is, of courst, married to Johansson.

2 My most-listened to voice is not 'Siri' nor 'Alexa', that would have to be 'Alex' the optional voice baked into macOS and iOS. Few people seem to know about it because "he" is buried in the accessibility settings and requires a seperate large download. But it's a really strong voice and has been for years!