The company announced Wednesday during its annual conference: MARS, which focuses on innovation in artificial intelligence, that it is working on an upgrade to its Alexa system that would allow the technology to mimic any voice, even a member. of the deceased family.
In a video shown on stage, Amazon (AMZN) demonstrated how, instead of Alexa’s signature voice reading a story to a child, it was her grandmother’s voice.
Rohit Prasad, senior vice president of Amazon, said the updated system will be able to collect enough voice data from less than a minute of audio to make personalization like this possible, instead of making someone spend hours in a recording studio as is done in the past. Prasad did not explain when this feature could be launched. Amazon declined to comment on a timeline.
The concept comes from Amazon looking for new ways to add more “human attributes” to artificial intelligence, especially “in these ongoing pandemic times, when many of us have lost someone we love,” Prasad said. “While AI can’t eliminate the pain of loss, it can definitely make your memories last.”
Amazon has long used recognizable voices, such as the real voices of Samuel L. Jackson, Melissa McCarthy and Shaquille O’Neal, to give Alexa a voice. But AI recreations of people’s voices have also improved more and more in recent years, especially with the use of AI and deepfake technology. For example, three lines of Anthony Bourdain’s documentary “Roadrunner” were generated by AI, even though they seemed to be told by the late media personality. (This particular case caused a stir because it was not made clear in the film that the dialogue was generated by AI and had not been approved by Bourdain’s property). “We can have a documentary ethics panel on this later,” director Morgan Neville told The New Yorker when the film premiered last year. More recently, actor Val Kilmer, who lost his voice due to throat cancer, partnered with startup Sonantic to create a Voice Speaking with Artificial Intelligence for him in the new film. ” Top Gun: Maverick “. The company used Kilmer’s archive audio footage to teach an algorithm how to speak like the actor, according to Variety.
Adam Wright, a senior analyst at IDC Research, said he sees the value of Amazon’s effort.
“I think Amazon is interested in doing that because it has the capacity and the technology, and they’re always looking for ways to elevate the smart assistant and the smart home experience,” Wright said. “It remains to be seen whether it drives a deeper connection with Alexa or simply becomes a skill that some people are entertained with from time to time.”
Amazon’s foray into Alexa’s personalized voices may struggle more with the strange valley effect: recreating a voice so similar to that of a loved one but not entirely correct, leading to the rejection of real humans .
“There are certainly some risks, as if the resulting voice and AI interactions don’t match well with the memories of that person’s loved ones,” said Micheal Inouye of ABI Research. “For some, they will see this as creepy or completely terrible, but for others it could be seen in a deeper way, like the example given by allowing a child to hear their grandparents’ voice, perhaps for the first time and in a way, it is not a strict record of the past “.
He believes, however, that the different reactions to ads like this talk about how society will have to adapt to the promise of innovations and its eventual reality in the coming years.
“We will definitely see more of this kind of experiments and trials, and at least until we get a higher level of comfort or these things become more common, there will still be a wide range of responses,” he said.