“Hi Alexa, what’s another word for ‘creepy’?”
In a demonstration comparing to the dystopian “Black Mirror” series, Amazon revealed that it has developed a way for its voice assistant Alexa to reproduce the speech of a dead relative, based on less than a minute of audio. engraving of the original person. .
The e-commerce giant unveiled the new technology on Wednesday at its re-MARS conference, Amazon’s global artificial intelligence event for machine learning, automation, robotics and space. In a video demonstration shown at the event, a child says, “Alexa, can Grandma finish reading ‘The Wizard of Oz’ to me?” – with which a synthesized voice of the grandmother emanates from an Amazon Echo Dot smart speaker.
“As you have seen in this experience, instead of Alexa’s voice reading the book, it is the voice of the child’s grandmother,” Rohit Prasad, senior vice president of Alexa AI and chief scientist, told attendees.
Amazon has already developed speech synthesis technology to allow Alexa to mimic the voices of celebrities such as Shaquille O’Neal and Melissa McCarthy. But this had previously required a person to record tens of hours of audio. Amazon has now developed a way to replicate a high-quality voice using less than a minute of recorded voice, which the company’s engineers were able to do “framing the problem as a voice conversion task and not as a task. of voice generation. ”according to Prasad.
Presenting the demonstration, Prasad said that Alexa’s AI ability to emulate the speech of a dead relative may represent a way to preserve the memory of a loved one who has died. “We certainly live in the golden age of AI, where our dreams and science fiction are becoming a reality,” he said.
Watch the demonstration of Alexa in the following video:
optional screen reader