Alexa might soon be able to mimic anyone’s voice
T-1000 in the making
Remember that scene from Terminator 2: Judgement Day where the T-1000 mimicked John Connor’s foster mother’s voice over the phone? Amazon’s digital assistant, Alexa, might soon be able to perform a similar feat if recent reports are to be believed.
Amazon is reportedly developing a new system (via Reuters) that will allow Alexa to mimic any voice. According to Rohit Prasad, a Senior Vice President at Amazon, the digital assistant will only need to hear less than a minute of audio to perform this amazing feat. Prasad further explained that the idea behind this new system was to give users the ability to “make the memories last” and listen to lost loved ones using Alexa-powered smart speakers.
Amazon showcased a demo video of the feature at a conference in Las Vegas yesterday. The video portrayed a child who asked, “Alexa, can grandma finish reading me the Wizard of Oz?” In response, Alexa affirmed the command and promptly changed its voice to a more soothing, less robotic version that apparently sounded like the child’s grandmother.
While the feature sounds impressive, it has already raised various concerns. Some believe that it could be used to create more realistic deepfakes, while others claim that it could be abused for impersonation and phone-based scams.
Interestingly, Amazon isn’t the only company working on such a feature. Last year, Microsoft opened limited access to a new text-to-speech AI called Custom Neural Voice. It’s part of an Azure AI service called Speech, and it uses multiple neural networks to mimic any voice or create a more natural-sounding synthetic voice. Several companies already use Microsoft’s tech to create unique experiences, like giving people the opportunity to interact with Bugs Bunny.
What’s your take on Alexa having the ability to mimic anyone’s voice? What would you use the feature for? Let us know in the comments section below.