Amazon is previewing the Alexa feature that synthesizes a person’s voice from less than a minute of audio

39

Rohit Prasad, Alexa AI Senior Vice President and Chief Scientist, talks about speech synthesis technology at Amazon re:MARS

LAS VEGAS – A short video of a kid asking Alexa to let her grandma read to her The Wizard of Oz generates more than its share of buzz after Amazon explained that the voice was actually a simulation.

To do that, Amazon had to “learn how to produce a high-quality voice in less than a minute of recording compared to hours of recording in a studio,” said Rohit Prasad, Alexa AI senior vice president and principal scientist, during a keynote address Wednesday morning the company’s re:MARS conference in Las Vegas.

Amazon doesn’t say when the feature might be released as an Alexa feature or what kinds of caveats might come with a rollout. AI-generated synthesized voices have raised concerns about the potential for abuse, which may be misused to trick people into believing others said something they didn’t say.

A lot of “Responsible AI” principles published by Microsoft this week outlines a number of safeguards against possible misuse of synthetic voices.

Hinting at the underlying technical approach used by Amazon for the capability, Prasad said that the key is “to define the problem as a language conversion task and not a language generation task”.

Presenting the demo, he described the technology as a way to preserve the memories of loved ones and said it’s part of a growing recognition of the role Alexa can serve as a companion.

“In this companion role, human attributes like empathy and affect are key to building trust,” he said. “These qualities have become even more important during these times of the ongoing pandemic when so many of us have lost someone we love. What AI can’t erase that pain of loss can definitely make their memories permanent.”

Prasad called it “one of the new skills we’re working on that will enable lasting personal relationships.”

Source link

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.