OpenAI’s recent foray into the world of voice technology has caused quite the stir. Their creation, “Sky,” a voice for ChatGPT, has been discovered to bear a striking resemblance to the vocal stylings of Hollywood actress Scarlett Johansson. A forensic analysis conducted by researchers at Arizona State University, commissioned by NPR, revealed that Sky’s voice shared uncanny similarities with Johansson’s, more so than 98 percent of other actresses evaluated in the study.
Interestingly, Johansson’s voice wasn’t always the frontrunner in the comparison. The analysis noted that voices of actresses like Anne Hathaway and Keri Russell were also frequently rated as akin to Sky. However, nuances such as the slight pitch difference and expressiveness set Johansson’s voice apart, with Sky leaning towards a higher pitch and more animated delivery compared to Johansson’s trademark breathy tones.
The controversy surrounding Sky escalated when it became apparent that the voice bore a striking resemblance to Johansson’s portrayal of an AI chatbot in the movie “Her.” OpenAI CEO Sam Altman’s tweet, simply stating “Her,” on the day of Sky’s release only added fuel to the fire, hinting at the inspiration behind the voice assistant. However, amidst mounting pressure and scrutiny, OpenAI swiftly retracted the Sky voice, denying any direct mimicry of Johansson’s voice.
Scarlett Johansson, on her part, revealed that OpenAI had approached her not once, but twice, seeking to license her voice for their project. Despite her refusal to lend her voice, Johansson was taken aback when she discovered the striking similarity between her voice and that of the contentious Sky. The debacle left both parties at odds, with Johansson feeling that her vocal identity had been exploited without consent.
In the realm of AI and voice technology, the incident serves as a cautionary tale about the ethical considerations and legal implications of leveraging real-life identities for virtual entities. The fine line between inspiration and imitation has been starkly drawn, prompting a reevaluation of the boundaries in voice technology development. As the dust settles on the Sky-Johansson debacle, it stands as a testament to the power of AI to blur the lines between reality and imitation, raising pertinent questions about consent, creativity, and accountability in the digital age.