OpenAI pauses ChatGPT voice that sounds like Scarlett Johansson


When OpenAI unveiled its new voice assistant technology this month, it immediately drew comparisons to Spike Jonze's 2013 sci-fi film “Her.”

Turns out the resemblance may have been too close for comfort.

OpenAI said Monday that it is pausing the use of one of the voices introduced as a new option for its artificial intelligence tool, ChatGPT. The San Francisco-based technology company said it was putting the voice known as “Sky” on pause after some people said it sounded similar to actress Scarlett Johansson's dulcet tones in the acclaimed film.

In “Her,” Johansson plays “Samantha,” the disembodied voice of a computer who provides friendship and, eventually, love to a lonely man played by Joaquin Phoenix.

“We've heard questions about how we choose voices on ChatGPT, especially Sky,” OpenAI posted on X on Monday. “We are working to suspend use of Sky while we address them.”

OpenAI CEO Sam Altman seemed to invite parallels to the movie in his announcement of the interactive voice feature, saying in a blog post that it “feels like movie AI.” “Her” was nominated for multiple Oscars, and Jonze won the Academy Award for original screenplay.

Despite the striking sonic resemblance, Johansson did not actually provide the voice for OpenAI's “Sky,” one of the multiple voice options available in the app. Rather, another actor was using his own “natural voice,” OpenAI said in a blog post on Monday.

“We believe that AI voices should not deliberately imitate a celebrity's distinctive voice,” the company said.

The move comes at a time when concerns have been raised within the entertainment industry over whether copyrighted material is used to train AI models.

OpenAI has said that its large language models, including those powering ChatGPT, are developed using publicly available information on the Internet, material acquired through licenses with third parties, and data provided by its users and “human trainers.”

The company has said it believes training AI models with publicly available materials on the Internet is “fair use.”

But some media outlets, including the New York Times, have sued OpenAI, concerned about how the tech company uses their stories.

Sony Music Group is in the process of sending hundreds of letters to AI developers and music streaming services, including OpenAI and Google, warning them not to use their artists' music to train generative AI tools without their permission. Actors and writers have also expressed concern about the impact of AI on their livelihoods.

Talent agencies are also being proactive in protecting their clients from unauthorized use of their image and voices. Century City-based Creative Artists Agency is helping clients through Vault, which scans clients' bodies and records their movements and voices to create a digital version of them to protect against copyright infringement.

Johansson is a CAA client. CAA did not immediately respond to a request for comment.

OpenAI said that ChatGPT's voices, named Breeze, Cove, Ember, Juniper and Sky, are from voice actors who went through an audition process. More than 400 people applied. Some of the characteristics OpenAI was looking for include “a voice that seems timeless” and “an approachable voice that inspires trust.”

The selected actors came to San Francisco to record their voices in June and July 2023, and their voices were added in September of that year, the company said. The company said the actors were compensated at “above-market rates” and that they were aware of the intentions and scope of the project.

“To protect your privacy, we cannot share the names of our speakers,” OpenAI said.

scroll to top