According to Amazon, by imitating the voice of anyone heard, Alexa will be able to “make memories eternal.”
The company is developing a technology that will allow its digital assistant, Alexa, to mimic the voice of anyone from an audio recording of less than a minute, chief scientist Rohitas Prasad said on Wednesday. He added that during the coronavirus pandemic, “so many of us have lost a loved one.”
Although no launch date has been specified, the core technology has been around for several years, writes The Guardian. The company staged a demonstration in which an older woman’s voice was used to read her granddaughter a tale before bed when she asked “Alexa, can my grandmother stop reading me“ The Wizard of Oz ”?
Apart from the initial demonstration, the details were few. The technology was announced at the company’s re: Mars conference, focusing on its “environmental computing” achievements in machine learning, automation, robotics, and space.
Praad added that the goal of Amazon’s voice assistant is “universal intelligence.”
However, other technology companies are wary of making digital voice secondary creation so easy: hours before Amazon announced its plans, Microsoft published new rules on the ethics of artificial intelligence that severely restrict who can create synthetic voices and how they can. could be used.
“It’s hard to imagine how this could be used to pretend to be a speaker and deceive listeners,” said Natasha Crampton, the company’s chief artificial intelligence specialist.
Microsoft will require companies to seek permission to create artificial voices, and began labeling them with an inaudible signal last month to detect misuse, according to The Guardian.