OpenAI Halts ChatGPT Voice Alleged to Imitate Scarlett Johansson

OpenAI, the company known for its advancements in artificial intelligence, has decided to put the usage of its new voice mode for ChatGPT on hold. This development comes in response to controversies regarding the selection and acquisition of one of the five voices, named “Sky.” On May 20, OpenAI communicated via social media platform X that it had received numerous questions and concerns about the Sky voice and was working to resolve them. The stir started after actress Scarlett Johansson made public statements about her interactions with OpenAI and its CEO, Sam Altman.

According to Johansson, Altman approached her with an offer to voice the ChatGPT 4.0 model, which she declined. Despite rejecting the proposal, Johansson was taken aback when friends and family informed her that the Sky voice sounded strikingly similar to her own. “I was shocked, angered, and in disbelief that Mr. Altman would choose a voice so eerily similar to mine,” Johansson stated. This prompted her to hire a legal team to address what she considers a serious issue in an era where “deepfakes” are becoming increasingly prevalent. She emphasized the importance of protecting one’s likeness and personal identity.

In its defense, OpenAI gave a thorough explanation of their casting and selection process for the ChatGPT voices, maintaining that Johansson was not intentionally imitated. The company clarified that Sky’s voice was performed by a professional actress using her natural speaking voice. OpenAI also noted that it believes AI voices should not deliberately mimic a celebrity’s unique vocal attributes. They provided each voice actor with a clear vision for the AI-human interaction, along with discussions about the technology’s potentials and limitations.

Sam Altman, the CEO of OpenAI, personally addressed the issue, asserting that any resemblance to Johansson’s voice was unintentional. His previous social media activity has not helped his case. On May 13, Altman posted “her” on platform X, which many internet users interpreted as a reference to the 2013 film titled “Her.” In the film, Johansson voices an AI system with which the main character falls in love. This context fueled speculation about whether the similarities were indeed accidental.

This situation is particularly significant for Scarlett Johansson, who has previously battled unauthorized use of her likeness in AI contexts. Seven months earlier, she filed a lawsuit against another AI company for similar reasons, marking her ongoing concern over her digital representation. The misuse of AI and deepfakes were notable issues during the SAG-AFTRA union strikes in Hollywood last year, where actors demanded regulations to protect their likenesses from being replicated without consent for new content.

Public figures have frequently been targeted by AI-generated impersonation scams over the past year. Personalities such as YouTube’s Mr. Beast, actor Tom Hanks, and entrepreneur Elon Musk have all experienced unauthorized AI impersonations. This rising trend has heightened public awareness and concern about the ethical implications of AI technology.

Scarlett Johansson isn’t the only one feeling harmed by unauthorized uses of their voice in AI systems. Take, for example, Susan Bennett, the original voice of Apple’s Siri assistant. Bennett recorded her voice in 2005 for ScanSoft, and it was later acquired by Apple for Siri without her knowledge. She expressed her disappointment in not receiving any compensation or acknowledgment from Apple for her voice that became widely recognized globally.

These incidents underscore a growing conflict at the intersection of technology and intellectual property. As AI continues to evolve and blend more seamlessly with human characteristics, the need for clear ethical guidelines and legal measures becomes more critical. The OpenAI controversy highlights the struggles public figures face in maintaining control over their own identity and likeness in the age of advanced artificial intelligence, bringing to light both the potentials and pitfalls of this rapidly developing tech.