Amazon Alexa will soon speak in the voice of your dead relative


Talking to friends and relatives from across the grave is an impossible task. But Amazon’s virtual assistant Alexa may soon do so. Whether you find that scary or comforting is entirely up to you.

At Amazon’s Re:MARS (Machine Learning, Automation, Robots and Space) conference, Alexa Senior Vice President Rohit Prasad demonstrated an early feature that Alexa might one day have: the ability to mimic voices. Equally, or perhaps more shocking, is the fact that this feature (or skill as they call it) enables Alexa to mimic the voices of people we’ve lost or are no longer with us.

Amazon also demonstrated this functionality in the form of a recorded video. In the demonstration video, a child says, “Alexa, can Grandma finish reading The Wizard of Oz to me?”. CNBC reported that upon hearing the request, Alexa accepts the child’s request in her normal voice, after which she begins to read the story in a voice that resembles that of the child’s deceased grandmother.

While it might sound borderline intimidating, the offerings at the company’s annual event introduced this functionality as a way to preserve memories. Amazon says Alexa’s ability to mimic people’s voices won’t “eliminate the pain of loss,” but it can certainly “make memories last.”

How does this feature work?

In case you’re curious how the feature works, Amazon told Engadget that Alexa’s new skill can create a synthetic voiceprint of a person’s voice after the individual voice is trained on one minute of audio. The advances the company has made in text-to-speech technology are what powers it. Amazon also recently shared a whitepaper detailing these developments, saying that a ‘voice filter’ could use as little as a minute of speech for Alexa to repeat the voice.

“State-of-the-art text-to-speech (TTS) systems require many hours of recorded speech data to generate high-quality synthetic speech … propose a TTS method that uses at least one minute of speech from the target speaker. It uses Voice Conversion (VC) as a post-processing module connected to an already existing high-quality TTS system and Marks a conceptual shift in the existing TTS paradigm, framing the few-shot TTS problem as a VC task. wrote in the white paper.

but there are concerns

Although this all sounds good enough, things are not as simple as they seem. Experts have long been concerned about the tools used to imitate voices in deep fake videos. Although this skill is still in development and it is unclear whether Amazon will release it to its users globally, it does raise concerns about the technology being misused by scammers and cybercriminals.

Amazon Alexa will soon speak in your dead relative’s voice The post first appeared on BGR India.

Read full article here

Leave a Reply