\n<\/p>\n
Researchers at Arizona State University found that Scarlett Johansson\u2019s voice does indeed sound like \u201cSky,\u201d OpenAI\u2019s currently discontinued<\/a> GPT-4o voice, according to NPR<\/em><\/a>, which commissioned the comparison. After using AI models developed to analyze vocal similarities to compare Sky to roughly 600 other actresses, the lab reportedly found Johansson\u2019s voice was \u201cmore similar to Sky than 98% of the other actresses.\u201d<\/p>\n<\/div>\n But the models also \u201coften\u201d found that the voices of Anne Hathaway and Keri Russell resembled Sky\u2019s more than did Johansson\u2019s. Visar Berisha, the professor who led the analysis, told NPR<\/em> that Johansson\u2019s voice is \u201csimilar, but likely not identical.\u201d Berisha\u2019s other work includes OriginStory<\/a> (PDF<\/a>), an FTC challenge-winning microphone<\/a> that watermarks voice recordings as human-created. <\/p>\n<\/div>\n Nevertheless, the researchers reportedly concluded there were \u201cundeniable commonalities\u201d between the voices. For instance, the analysis reportedly showed that Sky\u2019s and Johansson\u2019s vocal tracts would have \u201cidentical\u201d lengths (\u201cvocal tract\u201d refers to the throat, mouth, and nasal passages that would produce a given voice). <\/p>\n<\/div>\n Some differences included that Sky\u2019s voice was a little higher-pitched and more expressive than Johansson\u2019s, while hers is \u201cslightly more breathy\u201d than that of the model, according to NPR<\/em>. We asked Berisha for more information on his analysis and its limitations, and will update if he responds. OpenAI did not immediately respond to a request for comment.<\/p>\n<\/div>\n Both OpenAI CEO Sam Altman<\/a> and CTO Mira Murati<\/a> have denied that Sky is intended to sound like Johansson. After its GPT-4o demo earlier this month \u2014 after which Altman published a one-word post<\/a> that simply read \u201cher\u201d \u2014 Johansson said<\/a> that Altman had asked her to lend her voice to the model, which she declined, and that he tried again as recently as two days prior to the demo. <\/p>\n<\/div>\n