After rejecting an offer to lend his voice to a ChatGPT update, he pointed out that the company was using one very similar to his own.
It was mid-May 2024 when OpenAI released an update to its chatbot ChatGPT-4o which featured a female voice who interacted with users.
But beyond the technological capabilities of the system, one of the points that has attracted the most attention is that offered a tone similar to that of Scarlett Johansson in the movie Her (2013).
For the production of the said science fiction film, the American actress lent his voice to the character of a virtual assistant with whom the protagonist falls in love.
From the company of artificial intelligence (IA) had contacted Johansson to ask her to make their chatbot, but she declined to the company’s offer.
However, the similarities between her voice and that of the system led the actress to express publicly express their annoyance .
In a letter he shared, he said that She felt “surprised, angry and couldn’t believe” the relationship even if he had rejected the proposal.
And although OpenAI has maintained that the voice was not inspired by his own and was recorded by someone else, They stopped using it and regretted “not having communicated better” .
During the presentation of the new version of the chatbot, the general director of the firm, Sam Altman had written on his X account (ex Twitter): “ Her » .
This reference to the film Johansson participated in was interpreted by many users as actually being inspired by his voice .
The situation led many Internet users to express their criticism and the debate on regulation intensified what the development of AI should have, a field that has distinguished itself by its accelerated progress in recent years.
What Scarlett Johansson said about the voice used by OpenAI
In a letter recovered by New York Times On May 20, 2024, Johansson explained that “last September, I got an offer from Sam Altman, who wanted to hire me to voice him to the current system.
“He told me he thought that by giving the system a voice, he could bridge the gap between tech companies and creatives and help consumers feel comfortable with the radical change happening around humans and AI. He said he thought my voice would comfort people. After much thought and for personal reasons, I declined the offer.
However, the actress accused, “Nine months later, my friends, family and the general public noticed how much the new system called ‘Sky’ resembled me.” .
“When I heard the demo come out, I was shocked, angry, and couldn’t believe that Mr. Altman would seek out a voice that sounded so uncannily like mine. that my closest friends and the media couldn’t tell the difference.”
“Mr. Altman He even hinted that the similarity was intentional, tweeting a single word: “ Her ‘, a reference to the film in which I lent my voice to a chat system, Samantha, who establishes an intimate relationship with a human.
According to Johansson, two days before the demo was released, the company’s CEO AI He contacted his agent to ask him to “reconsider” .
“Before we could connect, the system was already available. As a result of his actions, I was forced to hire legal counsel, who wrote two letters to Mr. Altman and OpenAI, explaining what they had done and asking them to detail the process. “Exactly how they created the voice of ‘Sky’.”
“Therefore, OpenAI reluctantly agreed to remove voice of ‘Sky’,” the actress added.
Under this line, he said: “At a time when we are all faced with deepfakes and the protection of our own image, our own work, our own identity, I think these are issues that deserve absolute clarity. I look forward to a resolution in the form of transparency and the adoption of appropriate legislation to help ensure the protection of individual rights. “.
To which Sam Altman, the CEO of the AI company, responded
On May 20, 2024, Altman shared a statement in which He said that “Sky’s voice is not Scarlett Johansson’s and we never intended for it to sound like her.” .
After clarifying that they were specifically reaching out to someone else for the task, he announced that “Out of respect for Ms. Johansson, we have stopped using Sky’s voice in our products.” .
“We regret not having communicated better” he condemned.
Two days later, the company updated its official website an article in which They detailed “how the ChatGPT voices were chosen” .
“We worked with leading casting and directing professionals to narrow down over 400 submissions before selecting the top 5 voices.” synthesized from OpenAI in their note.
Doubts about copyright in the development of this technology
The episode that occurred between a celebrity like Johansson and the company headed by Altman was positioned as one of those with the greatest impact, but it is not the only one in which the mechanisms have been questioned that the company uses to develop its products.
Previously, the American actress and writer Sarah Silverman had accused the company of training its model with her memoirs titled The one who wets the bed (Vivace, 2011).
Other authors such as George RR Martin and John Grisham also joined a similar lawsuit. accused of resorting to “systematic thefts on a large scale”, according to comments obtained by the magazine time .
For his part, The New York Times presented one at the end of 2023 in which they assured that OpenAI was using their content to deliver it to their systems.
Regarding this matter, the firm shared an article on your website where They responded that the American newspaper was “not telling the whole story.” “.
At the same time, they stressed that Its goals are to “support a healthy information ecosystem, be a good partner and create opportunities.” mutually beneficial.
As saved Tutor OpenAI declared in the UK House of Lords in late 2023 that “given that copyright today covers virtually every type of human expression – including blogs, photographs, forum posts, bits of software code and government documents”, It would be impossible to train today’s leading AI models without using copyrighted material “.
A survey published in November of the same year by the Pew Research Center stated that, In the United States, 67% of people familiar with chatbots like ChatGPT expressed concern that the government is not going far enough on regulation. of its use.
On the other hand, 31% of the sample expressed the opposite .
Source: Latercera
I am David Jack and I have been working in the news industry for over 10 years. As an experienced journalist, I specialize in covering sports news with a focus on golf. My articles have been published by some of the most respected publications in the world including The New York Times and Sports Illustrated.
