Last summer, while on their way to a medical appointment near their home in New York, Paul Skye Lehrman and Linnea Sage listened to a podcast (podcast) about the rise of artificial intelligence (AI) and the threat it posed to the livelihoods of writers, actors and other entertainment professionals.
The subject was particularly important to this young married couple. Both made their living as voice actors, and artificial intelligence technologies were beginning to generate voices that sounded like the real thing.
But the podcast took an unexpected turn. To highlight the threat posed by AI, the host conducted a long interview with a conversational agent named Poe. This one looked exactly like Lehrman.
“He was questioning my voice about the dangers of AI and the harmful effects it could have on the entertainment industry,” explains Mr. Lehrman. We stopped the car and stood there in complete disbelief, trying to figure out what had just happened and what we should do. »
M. Lehrman and M.me Sage is now suing the company that created the robot’s voice. They claim that Lovo, a start-up of Berkeley, California, illegally used recordings of their voices to create technology that could rival their voice work. After hearing a clone of Mr. Lehrman’s voice on the podcast, the couple discovered that Lovo had done the same for Mr. Lehrman’s voice.me Wise.
PHOTO ELIANEL CLINTON, THE NEW YORK TIMES
Linnea Sage and Paul Lehrman
The couple joins a growing number of artists, editors, computer programmers and other creators who have sued makers of artificial intelligence technology, arguing that the companies used their work without permission to create tools that could ultimately replace them in the labor market. (In December, the New York Times filed a lawsuit against two of these companies, OpenAI and its partner Microsoft, accusing them of using its copyrighted news articles to create their online chat bots.)
In their complaint, filed Thursday in Manhattan federal court, the couple claims that anonymous Lovo employees paid them for a few voice clips in 2019 and 2020 without disclosing how they were used.
M. Lehrman and M.me Sage claim that Lovo, which was founded in 2019, violates federal trademark law and several state privacy laws by promoting clones of their voices. They invite other voice actors to join them.
“We don’t know how many other people were affected,” said their lawyer, Steve Cohen.
Lovo denies the allegations in the complaint, said David Case, an attorney representing the company. He added that if everyone who provided voice recordings to Lovo gave consent, “there is no problem.”
Tom Lee, the company’s CEO, said in a podcast episode last year that Lovo now offered a revenue-sharing program that allowed voice actors to help the company create voice clones of themselves- same and receive a share of the revenue generated by these clones.
“Test scripts”
The lawsuit appears to be the first of its kind, according to Jeffrey Bennett, general counsel for the Screen Actors Guild-American Federation of Television and Radio Artists, the union that represents 160,000 media workers worldwide.
This lawsuit will show people – especially tech companies – that there are rights to your voice, that there is a whole group of people who make a living using their voice.
Jeffrey Bennett, general counsel of the Screen Actors Guild-American Federation of Television and Radio Artists
In 2019, Mr. Lehrman and Mr.me Sage became known as voice actors on Fiverr, a website where independent professionals can advertise their work. Through this online marketplace, they were often asked to provide voices for commercials, radio spots, online videos, video games, and other media.
That year, Mme Sage was contacted by an anonymous person who paid her $400 to record several radio scripts and explained that the recordings would not be used for public purposes, according to correspondence cited in the complaint.
“These are test scripts for radio commercials,” the anonymous person said, according to the complaint. “They will not be disclosed externally and will only be used internally, so they will not require any rights of any kind. »
Seven months later, another unidentified person contacted Mr. Lehrman about a similar job. Mr. Lehrman, who also works as a television and film actor, asked how the clips would be used. The person responded multiple times that they would only be used for research and teaching purposes, according to correspondence cited in the lawsuit. Mr. Lehrman was paid $1,200. (He provided longer recordings than those of Mme Wise.)
In April 2022, Mr. Lehrman discovered a YouTube video about the war in Ukraine narrated by a voice that sounded like his own.
“It is my voice that speaks about armament in this conflict,” he said.
I turned white as a sheet – I got goosebumps. I knew I had never said those words in that order.
Paul Skye Lehrman, voice actor
For months, Mme He and Sage struggled to understand what had happened. They hired a lawyer to help them track down who made the YouTube video and how Mr. Lehrman’s voice was recreated. But the owner of the YouTube channel appeared to be based in Indonesia, and they had no way of tracking him down.
That’s when they heard the podcast on their way to the doctor’s office. Thanks to the podcast Deadline Strike Talk, they were able to identify the source of Mr. Lehrman’s voice clone. A professor at the Massachusetts Institute of Technology assembled the conversational robot using Lovo’s text-to-speech technology.
Accessible voice cloning
M. Lehrman and M.me Sage also discovered that Lovo was promoting voice clones of them on his website. After they sent the company a formal notice, the company said it had removed their voice clones from the site. But they argued that the software that drove these voice clones had already been downloaded by countless numbers of the company’s customers and could still be used.
Mr. Lehrman also questioned whether the company used the couple’s voices, along with many others, to develop the core technology for its voice cloning system. Speech synthesizers often learn their skills by analyzing thousands of hours of speech, in the same way that OpenAI’s ChatGPT and other conversational bots learn their skills by analyzing large amounts of text taken from the internet.
Lovo admitted to training his technology using thousands of hours of recordings of thousands of voices, according to correspondence filed in the lawsuit.
Mr. Case, the lawyer representing Lovo, said the company trained its artificial intelligence system using audio recordings from a freely available database of English-language recordings called Openslr.org . He did not answer a question about whether the recordings of the voices of Mr. Lehrman and Ms.me Sage had been used to train the technology.
This article was published in the New York Times.
Read the original version (in English; subscription required)