Scarlett Johansson has expressed strong disapproval of OpenAI and its founder Sam Altman for creating a chatbot voice she claims sounds “eerily similar” to her own. Johansson had previously declined to license her voice to the AI 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,” Johansson stated in a message published by NPR on Monday.
Introduction of GPT-4 Omni
Last week, OpenAI unveiled GPT-4 Omni (GPT-4o), its latest AI model. The company showcased how the AI can engage in more human-like conversations, complete with abilities to whisper, make sarcastic remarks, and even flirt. This demonstration led to comparisons with Johansson’s character Samantha from the 2013 film Her, directed by Spike Jonze, where Johansson voiced a virtual assistant that forms an intimate relationship with a lonely writer.
Suspension of AI voice
On Sunday, OpenAI announced it had suspended the use of its AI voice, Sky, to address concerns regarding the similarity to Johansson’s voice. Sky has been available since the launch of ChatGPT’s voice mode last September and was one of five voices featured with GPT-4o. In a blog post, OpenAI emphasized that it did not copy Johansson’s voice.
“We believe that AI voices should not deliberately mimic a celebrity’s distinctive voice — Sky’s voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice,” the company explained. “To protect their privacy, we cannot share the names of our voice talents.”
Altman’s response
Sam Altman reiterated OpenAI’s position on Monday. In a statement, Altman clarified, “The voice of Sky was never intended to resemble the voice of Johansson. We cast the voice actor behind Sky’s voice before any outreach to Ms. Johansson. Out of respect for Ms. Johansson, we have paused using Sky’s voice in our products. We are sorry to Ms. Johansson that we didn’t communicate better.”
Johansson’s accusations
In her statement, Johansson accused the company and Altman of intentionally copying her voice. She revealed that Altman had approached her in September to voice GPT-4o, suggesting her voice would help consumers “feel comfortable with the seismic shift concerning humans and AI.” Despite declining the initial offer for personal reasons, Johansson noted that Altman approached her agent again shortly before the May event, urging her to reconsider licensing her voice for a virtual assistant.
“Before we could connect, the system was out there,” Johansson said, adding that she was “forced” to hire legal counsel. Her lawyers have written to Altman, demanding transparency regarding the process of hiring the voice talent.
Voice selection process
OpenAI’s blog post detailed the selection process for the AI voice, stating that the company narrowed down over 400 submissions from voice and screen actors to five. These chosen voices were selected based on characteristics such as “an approachable voice that inspires trust,” “feels timeless,” and “is natural and easy to listen to.”
Broader legal challenges
Johansson’s legal threat comes amidst a series of copyright disputes facing OpenAI. The company is grappling with allegations from various creative industries, including Hollywood and the broader media sector. In April, a group of eight daily newspapers filed a lawsuit against OpenAI and Microsoft, accusing them of copyright infringement related to the unauthorized use of their articles for training AI models.
As OpenAI continues to navigate these legal challenges, Johansson’s criticism highlights the ongoing complexities and ethical considerations in the development and deployment of AI technologies.
