Actress Scarlett Johansson said she was “shocked” and “angered” Monday after ChatGPT rolled out its voiced artificial intelligence (AI) assistant that she claims sounds “eerily similar” to her voice.
Johansson, in a statement, noted that OpenAI CEO Sam Altman previously reached out to her about voicing an AI for the company, but she declined. The company released a demo for its assistant with the “Sky” voice last week.
“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,” she said in a statement to The Hill.
“Two days before the ChatGPT 4.0 demo was released, Mr. Altman contacted my agent, asking me to reconsider,” she continued. “Before we could connect, the system was out there.”
Altman appeared to acknowledge the similarity, posting a single word — “her” — on social media platform X, possibly in reference to the 2013 film of the same title, in which Johansson plays an artificial intelligence that has a relationship with a person.
Johansson said she hired legal counsel to request OpenAI take down the “Sky” voice, which it “reluctantly” agreed to do.
“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,” she said. “I look forward to resolution in the form of transparency and the passage of appropriate legislation to help ensure that individual rights are protected.”
OpenAI replied in a statement Monday, saying the “Sky” voice is not a mimic of the Hollywood star.
“We believe that AI voices should not deliberately mimic a celebrity’s distinctive voice," the company wrote. "Sky’s voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice."
The controversy was also featured on “Saturday Night Live” over the weekend, with “Weekend Update” co-host Colin Jost — Johansson’s husband — forced to read a joke penned by his co-host poking fun at his wife.
“ChatGPT has released a new voice assistant feature inspired by Scarlett Johansson’s AI character in ‘Her,’ which I’ve never bothered to watch because without that body, what’s the point of listening?” he said.
Members of Congress have raised concerns over deepfakes and AI voice and image mimics in recent months, with the issue garnering its own Senate Judiciary Committee hearing last month.
Singer FKA Twigs said in the hearing that she developed her own deepfake voice which sounded like herself to better engage with fans globally, but unauthorized songs featuring mimic voices make her and other artists feel “vulnerable.”
“Ultimately, what it boils down to is my spirit, my art, and my brand is my brand and I’ve spent years developing it and it’s mine,” she wrote in written testimony. “It doesn’t belong to anybody else for — to be used in a commercial sense or cultural sense or even just for a laugh."
"You know, I am me. I am a human being and we have to protect that," she added.
Similar deepfake technology has also been used for crime. Late last month, a deepfake voice impersonated a Maryland school principal, creating a fake recording which included racist and antisemitic remarks. The controversy forced the principal on leave before it was discovered to be manufactured.
Concerns around AI's potential impact on the upcoming election have also been raised. The Federal Communications Commission (FCC) was brought into the conversation earlier this year after a digitally altered robocall message created to sound like President Biden urged New Hampshire residents not to vote in February's primary.
Judy Kurtz contributed reporting.
