OpenAI used a voice that sounded a lot like the actress Scarlett Johansson's as its new ChatGPT voice assistant, as many pointed out after its debut.

Now things just got a whole lot more explosive, with Johansson saying that company's leadership asked her permission to use her voice — and then went ahead and did it anyway when she said no.

Johansson said in a new statement that OpenAI's Sam Altman wanted to hire her to voice the AI assistant, known as "Sky," which is one of five voice options available in the latest update of ChatGPT and was featured prominently in promotional videos.

The "Lost in Translation" star ultimately declined, and was appalled when she heard how much the finished product sounded like her own voice.

"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 said.

Worse, it feels almost indisputable that the company is explicitly cribbing her performance in the 2013 romantic tragedy "Her," with Altman tweeting the single word "her" as the chatbot launched — a move, in Johansson's reckoning, that  "insinuated that the similarity was intentional."

Just before the ChatGPT 4o launch, the actress wrote, the CEO contacted her agent asking her to reconsider her endorsement of Sky's voice. But before she had a chance to respond, "the system was out there."

"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," Johansson continued. "Consequently, OpenAI reluctantly agreed to take down the 'Sky' voice."

The company's response isn't doing anything to make it look less guilty. Before Johansson went public, OpenAI paused the Sky voice option and released a blog post about its selection process, claiming that hundreds of submissions were used to create the five voice options. In the post, OpenAI denied that Johansson's voice was used and said it instead belonged to a different voice actress.

Then, after Johansson went after the company, Altman released his own statement saying that Sky had been paused "out of respect" for the star.

The fallout for OpenAI may well extend beyond a bungled PR move illustrating its "ask for forgiveness, not permission" ethos.

The courts have often favored celebrities whose likenesses have been deployed without permission, and the fact that it explicitly reached out to Johansson — nevermind Altman's coy nod to the ripping off on social media — won't help its case.

And Johansson has a history of successful litigation: a few years back, she took on the entertainment behemoth Disney over a compensation dispute and forced a settlement.

Maybe the most significant lesson, though, is the one that the public will take: that from its leadership on down, OpenAI believes it can seize anyone's intellectual property or likeness and do whatever it wants with it — permission be damned.

"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 wrote. "I look forward to resolution in the form of transparency and the passage of appropriate legislation to help ensure that individual rights are protected."

More on OpenAI: The New ChatGPT Has a Huge Problem in Chinese

Share This Article