By learning from just a minute of audio, the new feature can give Amazon’s Alexa the ability to sound like virtually anyone.
Technology can’t bring back people from the dead just yet, but Amazon is working on a feature that will allow its AI assistant Alexa to speak in the voice of a deceased relative to “make the memories last”.
At the company’s Re:MARS AI conference in Las Vegas yesterday (22 June), senior vice-president at Amazon and head scientist for Alexa, Rohit Prasad, said it is working on technology that will give Alexa the ability to mimic virtually anyone’s voice based on just one minute’s audio.
Pitching it as a feature to help people cope with the loss of a loved one in light of the pandemic, Prasad showed the audience a video in which a child asks Alexa if it can read The Wizard of Oz in their grandmother’s voice. After briefly acknowledging the child in its own voice, Alexa switched to what appeared to be the grandmother’s voice and read the story to the child.
“This required inventions where we had to learn to produce a high-quality voice with less than a minute of recording versus hours of recording the studio,” Prasad said of the feature.
“The way we made it happen is by framing the problem as a voice-conversion task and not a speech-generation path. We are unquestionably living in the golden era of AI, where our dreams and science fictions are becoming a reality.”
Voice impersonation concerns
It is not clear yet when the feature could be available for roll-out to the public, but the technology no doubt raises some eyebrows for its uses beyond just mimicking the voice of a dead person. For one, it could be used to deepfake the voice of living people for malicious purposes.
“The phone attacking implications of this tool are not good at all – this will likely be used for impersonation,” tweeted hacker and CEO of SocialProof Security, Rachel Tobac.
“I know this technology already exists. I’ve talked about this risk with other orgs tools. But the easier it is to use, the more it will be abused. And this sounds like it may be pretty user friendly.”
Tobac argued in a Twitter thread that in addition to being abused for impersonation in calls and account takeovers, the technology could be used “to harass and harm individuals, especially the ones most vulnerable”.
“I look forward to Alexa using my dead father’s voice to tell me how proud he is of my brother,” joked another Twitter user and former journalist, Mike Sharp.
The AI voice technology space is seeing a lot of interest from big players. Streaming giant Spotify recently snapped up Irish-founded AI platform Sonantic, the company that gave life to the voice of Val Kilmer in Top Gun: Maverick.
At Silicon Republic’s Future Human event last month, Marino Software CEO Keith Davey and Maynooth University’s Trevor Vaugh spoke about how voice-banking tech can change the lives of people with motor neurone disease who lose their voice.
But they also flagged the potential ethical issues surrounding the use of voice technology after someone has died, and said it is important that people have a way to control what happens to their voice after their death so there is no risk of abuse.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.