At its re:MARS 2022 developer convention, Amazon revealed a novel experimental function for its voice assistant Alexa — the flexibility to copy the voice of deceased folks. Sure, you learn that proper. In a improvement that sounds straight out of a dystopian sci-fi film, Alexa might quickly have the choice to imitate your useless family.
Alexa Might Quickly Copy Voice of Deceased Family
On the occasion, Amazon demonstrated an instance video the place a baby asks Alexa to learn a bedtime story. “Alexa, Can grandma end studying me The Wizard of Oz?,” says the child within the video. Quickly after, Alexa begins reciting the story in what’s presumably the kid’s grandmother’s voice.
One of the best half is that the grandma’s voice sounded pure and never one thing you’ll count on from a voice assistant that normally sounds mundane and robotic.
“This required innovations the place we needed to be taught to provide a high-quality voice with lower than a minute of the recording versus hours of recording within the studio. The best way we made it occur is by framing the issue as a voice conversion activity and never a speech era path. We’re unquestionably residing within the golden period of AI, the place our desires and science fiction have gotten a actuality,” stated Rohit Prasad, Amazon’s Senior Vice President and Head Scientist for Alexa.
Whereas Amazon’s demonstration of the voice mimicking function is heartwarming, I can already consider varied circumstances of customers attempting to abuse the function. It stays to be seen if different voice assistants will intelligently detect that the voice is popping out of Alexa and keep away from circumstances of bypassing voice recognition authentication methods.
Prasad has additionally talked concerning the doable challenges as that is extra of a ‘voice conversion activity and never a speech era path.‘ We will count on extra particulars on this within the close to future. Relating to the provision, there’s no phrase for now as it’s at present below improvement. So, do you suppose AIs ought to have the choice to imitate the voices of useless folks? Share your ideas with us within the feedback.