Amazon’s Alexa is almost like a family member to many, with people increasingly becoming dependent on it for everyday tasks. Now the company is working to give the machine a new feature that could sound exciting to some but eerie to others. At its annual re:MARS conference on 22 June, Amazon revealed that Alexa could soon be able to recreate anyone’s voice — even if they are dead.
The re:MARS conference is Amazon’s annual global artificial intelligence event where the e-commerce giant showcases its progress with machine learning, automation, robotics and space. The voice replication feature was shown in a video demo to attendees at the conference held in Las Vegas.
More about Amazon Alexa’s AI voice replication tech
Demo shows Amazon Echo Dot speaking in the voice of a boy’s grandma
Rohit Prasad, Amazon’s Alexa AI senior vice president and head scientist, presented the demo. According to Prasad, Alexa will be able to recreate the voice of a dead relative based on less than a minute of audio recording of the original person.
“While AI can’t eliminate that pain of loss, it can definitely make their memories last,” he said, before the demo was shown to the attendees.
In the demo, a young boy asks Amazon Echo Dot, “Alexa, can Grandma finish reading me ‘The Wizard of Oz’?”
The Amazon Echo Dot responds with an “Ok” and then speaks in the voice of the boy’s grandmother.
“As you saw in this experience, instead of Alexa’s voice reading the book, it’s the kid’s grandma’s voice,” Prasad told attendees.
“We are unquestionably living in the golden era of AI, where our dreams and science fiction are becoming a reality,” he added.
More advanced than existing voice tech
The new voice replication feature is in development. Neither Prasad nor Amazon has yet revealed the date of its public launch.
Amazon already has a voice-synthesizing feature that lets Alexa mimic voices of celebrities such as Shaquille O’Neal and Samuel L. Jackson and even fictional character voices such as Marvel’s Deadpool and Star Wars’ R2-D2.
But till now the feature required dozens of hours of audio recording to be able to replicate the voice. The new feature, on the contrary, requires a few seconds of original voice recording.
Amazon’s latest development has triggered a new debate around AI technology, especially since it comes close on the heels of the claim by a Google engineer that the search engine giant’s chatbot named LaMDA has gained sentience.
(Main image: Andres Urena/@andresurena/Unsplash; Featured image: Nicolas J Leclercq/@nicolasjleclercq/Unsplash)