File this within the “sure that’s very intelligent of you, however please don’t” pile.
Amazon has demonstrated an experimental function that demonstrates how a toddler can select to have a bedside story learn to him by his Alexa… utilizing his lifeless grandmother’s voice.
Rohit Prasad, Amazon’s head scientist for Alexa AI, instructed attendees of its annual MARS convention that:
“…in these instances of the continuing pandemic, so many people have misplaced somebody we love. Whereas AI can’t eradicate that ache of loss, it might probably undoubtedly make their reminiscences final.”
Amazon says that its AI programs can learn to mimic somebody’s voice from only a single minute’s price of recorded audio.
Which appears a little bit creepy to me. As a result of likelihood is that many people have far more of our voices than that recorded someplace – whether or not it’s on voicemails, movies, or – oops! – podcasts.
And your voice could not simply be used to console little Benny at bedtime. It may additionally be abused to unlock your smartphone or to converse to HMRC.
Fortunately there isn’t any suggestion (but) that Amazon goes to launch this performance to the broader world. However give them time, give them time.
Amazon is way from the one firm with the smarts to fairly convincingly mimic somebody’s voice from only a small snatch of audio, however that doesn’t imply it’s a cool factor to do. And there are so so some ways by which it might be abused…
So, what’s the answer? How can we cease folks utilizing deepfaked variations of our voice with out our permission?
I’m unsure we will. Perhaps it might be cool if the boffins at Amazon thought of how you can resolve that downside as an alternative of educating Alexa to learn “The Wizard of Oz” utilizing the voice of a lifeless girl.
Discovered this text attention-grabbing? Observe Graham Cluley on Twitter to learn extra of the unique content material we put up.
