Alexa can already speak in a variety of different voices, from its new male option to celebrities such as Melissa McCarthy and Shaquille O’Neal. But researchers at Amazon are looking to take Alexa’s speech abilities to the next level by letting her mimic the voices of loved ones, even from beyond the grave.
During a presentation at Amazon’s re:MARS event this week, Alexa head scientist Rohit Prasad showed how Alexa could speak in the voice of a child’s (presumably deceased) grandmother, with the voice assistant reading a passage from The Wizard of Oz to the delighted young boy.
Prasad framed the ability as a way to ease the loss of a loved one by “mak[ing] their memories last,” adding that his team had found a way for Alexa to “produce a high-quality voice” using less than a minute of recording.
Prasad didn’t announce when–or even if–the new Alexa voice feature would be released to the general public, saying only that his team was “working” on the project.
The new Alexa ability is certainly impressive, but also plenty creepy. The idea of hearing an AI-powered voice of a long departed relative is as unsettling as it is comforting, and if Alexa could mimic a relative after listening to their voice for less than a minute, who else could she mimic?
Of course, the deepfake genie is already well out of the bottle, so perhaps Amazon can put it to good use with Alexa’s help. And indeed, Prasad went for a more utopian approach while demonstrating the feature, saying that we are “living in the golden era of AI where our dreams and science fiction are becoming a reality,”
Maybe so! But forgive those of us who are still a tad creeped out.
Ben has been writing about technology and consumer electronics for more than 20 years. A PCWorld contributor since 2014, Ben joined TechHive in 2019, where he covers smart speakers, soundbars, and other smart and home-theater devices. You can follow Ben on Twitter.