Actor Scarlett Johansson’s odd tussle with OpenAI over the voice used for “Sky,” its GPT-40 chatbot, may get airtime in Congress after a House subcommittee invited her to testify about it.
In a letter sent Monday, Rep. Nancy Mace, R-S.C., asked Johansson to speak at a hearing about deepfake technology held by the House Oversight’s Subcommittee on Cybersecurity, Information Technology and Government Innovation on July 9.
“You recently expressed concerns via social media about the resemblance between your own voice and that of the ‘Sky’ chatbot, recently released as part of OpenAI’s GPT-40 update,” Mace wrote. “This hearing would provide a platform for you to share those concerns with House Members, and to inform the broader public debate concerning deepfakes.”
According to Axios, Johansson’s team told Mace’s office that she cannot appear at the July 9 hearing but may be available in October.
Last month, OpenAI unveiled the latest update to its chatbot with a voice that sounded strikingly similar to Johansson’s. (My colleague Zeeshan Aleem wrote about the apparently deliberate similarities between GPT-40 and the AI character that Johansson voiced in the 2013 sci-fi movie, “Her.”) Days after “Sky” was released, Johansson released a statement saying that she had turned down an offer last year from OpenAI’s CEO, Sam Altman, to lend her voice to the chatbot.
“Nine months later, my friends, family and the general public all noted how much the newest system named ‘Sky’ sounded like me,” she wrote. “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 that my closest friends and news outlets could not tell the difference.”
Altman responded in a statement that the chatbot’s voice “was never intended to resemble” Johansson’s, and The Washington Post later confirmed his claim that the company hired an actress to voice “Sky” months before Altman reached out to Johansson.
Still, the controversy raised further questions about AI’s potential encroachment on culture and how it may affirm some of our biases. It also spurred wide discussion about AI deepfakes, which are increasingly being used to spread fake sexually explicit videos of female celebrities — and teen girls.
