‘Harry Potter’ narrator Stephen Fry says AI was used to steal his voice

Stephen Fry, the British actor behind the Harry Potter audiobook series, has revealed that artificial intelligence was used to imitate his unique accent in a documentary he had no idea about.

“I never said anything like that, it was a machine. Yes, I was shocked,” said Frye, who narrated more than 100 hours of the popular J.K. Rowling series.

“They used my reading of the seven Harry Potter books and created an AI of my voice from that dataset to create a new narration.”

Fry played some of the AI-generated audio at last week’s CogX festival in London, warning people about the dangers posed by the technology, according to . luck.

“What you heard is not the result of a mashup. It’s a flexible synthetic voice that adjusts the words to fit the meaning of each sentence,” Fry said.

“So everything from calls to storm Congress to hard porn could be read without my knowledge and without my permission. And this is what you just heard. But it was done without my knowledge. So I heard about this and sent it to agents on both sides of the Atlantic, and they went ballistic – they did that. I didn’t know something like that was possible.”

Stephen Fry said Thursday at the CogX Festival in London that his voice is being used by AI technology to generate narration for historical documentaries. “It won’t be long before full deepfake videos are just as convincing,” he says.

The growing presence of AI is a major issue in the ongoing writers strike that has shut down Hollywood.

Fry did not say whether he would sue the unidentified company behind the AI, but said the audio clip was “not going to be long before full-blown deepfake videos will be just as convincing. ”, he warned.

“Technology is a verb, not a noun, and it’s always in motion. What we have now is not what we will be in the future. When it comes to AI models, what we have now is “It’s going to advance faster than any technology I’ve seen,” he said.

“There’s one thing we can all agree on: what a strange time to live in.”

Deepfake technology, which replaces one person in an image or video with another, is already being used to deceive people, according to a recent study.

last month, university of london created 50 deepfake audio samples in English and Chinese using a text-to-speech algorithm trained on two publicly available datasets.

The samples were played to 529 study participants to try to decipher the real voices from the fake voices.

Participants were able to identify the fake audio only 73% of the time (1 in 4 times). This only improved slightly after receiving training on how to recognize deepfake audio.

Additionally, despite Twitter/X’s policy against media and non-consensual nudity, deepfakes of X-rated celebrities are circulating online in large numbers.

Mr. Fry claimed that the AI-generated voices were generated in the following manner: "Read 7 volumes of Harry Potter books." It will be over 100 hours in total.
Fry (pictured with JK Rowling) claims that the AI-generated audio was generated using “reading all seven Harry Potter books,” totaling more than 100 hours. .
Getty Images

Most of these reality-bending videos are eventually removed from the platform, but not before racking up tens of millions of views.

One snippet that gained attention on X in June showed the face of 22-year-old influencer Addison Rae Easterling over the body of another woman lying seductively on a bed.

The tweet had been viewed more than 21 million times, and the thread contained even more explicit deepfakes before the account owner deleted it following a series of outrages.

Earlier this year, portraits of actresses Emma Roberts and Scarlett Johansson were also used in sexually suggestive ads for deepfake software, which were later removed by Facebook.

Meanwhile, a teacher in Texas was the victim of digitally altered nude photos.

Deepfake content is infiltrating politics, and as the threat of misinformation posed by exploding technology grows, Google has announced that it will require “all verified election advertisers” to use AI in their ads. If there is a legal requirement, it must be “disclosed conspicuously.”

Google’s decision comes as other major AI companies warn that the technology could be used to undermine elections.

Deepfakes are already in circulation, including this ad featuring actress Emma Roberts' likeness.
Deepfakes are already in circulation, including this ad featuring actress Emma Roberts’ likeness.
deep fake

This reality-altering technology has also been seen in political propaganda, such as this image depicting Donald Trump being arrested.
This reality-altering technology has also been seen in political propaganda, such as this image depicting Donald Trump being arrested.
Twitter / Elliot Higgins

Deepfake images of Donald Trump resisting arrest and his wife Melania screaming at police have already gone viral.

Another recently introduced deepfake is President Trump hugs Dr. Anthony Fauci A campaign ad created by Republican presidential candidate Ron DeSantis using AI during the coronavirus pandemic.

President Joe Biden has also been targeted by several AI-powered ads, including one in April after the Republican National Committee reelected him to a second term with Vice President Kamala Harris. It also includes an ad showing the president celebrating.

The 30-second video showed a cheerful Biden and Harris, then cut to a harrowing scene of China invading Taiwan and shutting down American banks and crime-infested cities.

Leave a Reply

Your email address will not be published. Required fields are marked *