Getty PhotosHollywood star Scarlett Johansson has mentioned she was left “shocked” and “angered” after OpenAI launched a chatbot with an “eerily related” voice to her personal.
The actress mentioned she had beforehand turned down an strategy by the corporate to voice its new chatbot, which reads textual content aloud to customers.
When the brand new mannequin debuted final week commentators have been fast to attract comparisons between the chatbot’s “Sky” voice and Johansson’s within the 2013 movie Her.
OpenAI mentioned on Monday that it will take away the voice, however insisted that it was not meant to be an “imitation” of the star.
Nevertheless, Johansson accused the corporate, and its founder Sam Altman, of intentionally copying her voice, in a press release seen by the BBC on Monday night.
“Once I heard the launched demo, I used to be shocked, angered and in disbelief that Mr Altman would pursue a voice that sounded so eerily much like mine,” she wrote.
“Mr Altman even insinuated that the similarity was intentional, tweeting a single phrase ‘her’ – a reference to the movie by which I voiced a chat system, Samantha, who types an intimate relationship with a human.”
Set within the close to future, 2013 movie Her sees Joaquin Phoenix fall in love together with his machine’s working system, which is voiced by Ms Johansson.
Enable Twitter content material?
The actress, who has been nominated for 2 Academy Awards, mentioned she had been initially approached by Mr Altman about voicing the brand new chatbot in September.
“[Mr Altman] instructed me that he felt that by my voicing the system, I might bridge the hole between tech firms and creatives and assist customers to really feel comfy with the seismic shift regarding people and AI,” Johansson wrote.
“He mentioned he felt that my voice could be comforting to folks.”
However she ultimately rejected the supply for private causes, she mentioned.
Two days earlier than the Sky chatbot was launched, she added, Mr Altman contacted her agent, urging Johansson to rethink her preliminary refusal to co-operate with the corporate.
Including that she had been pressured to rent attorneys, the actress mentioned she had despatched two authorized letters to the corporate, to determine how the voice had been made.
“In a time after we are all grappling with deepfakes and the safety of our personal likeness, our personal work, our personal identities, I imagine these are questions that deserve absolute readability,” she wrote.
In a press release shared with the BBC by OpenAI, Mr Altman denied that the corporate had sought to mimic Johansson’s voice.
“The voice of Sky just isn’t Scarlett Johansson’s, and it was by no means meant to resemble hers,” he wrote.
“We forged the voice actor behind Sky’s voice earlier than any outreach to Ms Johansson. Out of respect for Ms Johansson, we’ve paused utilizing Sky’s voice in our merchandise. We’re sorry to Ms Johansson that we didn’t talk higher.”
Individually, the agency mentioned it was “working to pause” the voice whereas it addressed questions on the way it was chosen in a submit on X, previously Twitter.
In its weblog submit, OpenAI mentioned the 5 voices utilized by its chatbot had been sampled from voice actors it partnered with.
Copyright challenges
It comes simply six months after actors agreed to finish strikes which froze the leisure business amid requires higher pay and safeguards on using AI.
Ms Johansson took half in final 12 months’s industrial motion, which was partly about how studios would use AI to mimic actors’ faces and voices.
The deal struck with studios included reassurances that these wouldn’t be used with out consent from actors.
“To make use of somebody’s voice with out permission feels notably invasive at a time when mistrust of AI and concern over its potential harms are rampant,” mentioned Dan Stein, head of AI voice licensing firm Voice-Swap.
“Whether or not OpenAI skilled their new Sky voice utilizing audio from Scarlett Johansson or a sound-a-like, the very fact stays that she refused permission and her identification was exploited regardless.
“It units a harmful precedent for copyright and consent if essentially the most distinguished firm within the discipline behaves on this approach.”
OpenAI has been battling varied authorized challenges to the way it makes use of copyrighted info accessible on-line.
In December, the New York Occasions mentioned it deliberate to launch a lawsuit towards the company over allegations that it had used “hundreds of thousands” of articles revealed by the media organisation to coach its ChatGPT AI mannequin.
And in September, authors George RR Martin and John Grisham additionally introduced a plan to pursue a declare, over allegations their copyright had been infringed to coach the system.
