Amazon shows software for instant voice copying
Amazon Echo speaker. The software should now be able to learn voices from people.
Amazon software is said to be able to imitate human voices. Artificial intelligence should only need a speech sample of less than a minute. But there is also criticism.
Updated: 06/24/2022, 15:09
Las Vegas. Amazon has developed software that can imitate a human voice in less than a minute.
The online group demonstrated on Thursday a network speaker with voice assistant Alexa reading the book “The Wizard of Oz” to a boy using his grandmother’s voice.
The online company raised concerns about the use of software to preserve the voices of deceased family members. Alexa’s research director Rohit Prasad said that many relatives died in the corona epidemic. “Artificial intelligence can’t remove that pain, but it certainly can keep the memory alive,” he said before demonstrating the technique. In an interview with the technology blog “TechCrunch”, Prasad then emphasized that Grandma made a living from the performance. “It wasn’t about ‘dead grandma’.”
Amazon’s technological breakthrough is that the software can imitate speech with so little raw data. In the past, you had to record texts for several hours. The technology is experimental and it’s not clear if it will reach consumers. Siwei Liu, a professor of computer science at the University of Buffalo, expressed concern about the possibility of the software being misused. In the financial services provider Bloomberg, he warned that criminals could pose as family members or someone with false statements from top managers could mess up the stock markets.
© dpa-infocom, dpa:220624-99-781130/3