LAS VEGAS — A brief video of a youngster asking Alexa to have his grandma learn the Wizard of Oz is producing extra than its share of buzz after Amazon defined that the voice was truly a simulation.
To make it occur, Amazon “needed to be taught to supply a high-quality voice with less than a minute of recording versus hours of recording in a studio,” mentioned Rohit Prasad, Alexa AI senior vp and head scientist, throughout a keynote handle Wednesday morning on the firm’s re:MARS convention in Las Vegas.
Amazon isn’t saying when the capability could possibly be launched as an Alexa characteristic, or what sorts of precautions may accompany a rollout. AI-generated synthesized voices have raised considerations concerning the potential for abuse, presumably being misused to trick individuals into considering others had mentioned one thing they hadn’t.
A set of “Accountable AI” ideas launched this week by Microsoft outlines a sequence of safeguards towards potential misuse of artificial voices.
Prasad hinted on the underlying technical method utilized by Amazon for the capability, saying that the important thing was “framing the issue as a voice-conversion job, and never a speech-generation job.”
Introducing the demo, he described the expertise as a option to protect the reminiscences of family members, saying it’s half of a rising recognition of the position that Alexa can function a companion.
“On this companionship position, human attributes of empathy and have an effect on are key for constructing belief,” he mentioned. “These attributes have change into much more necessary in these occasions of the continued pandemic, when so many of us have misplaced somebody we love. What AI can’t remove that ache of loss, it may possibly positively make their reminiscences final.”
Prasad known as it “one of the brand new capabilities we’re engaged on, which permits lasting private relationships.”
Credit score: Source link