Do you miss the sound of a dead relative’s voice?
Fear not: Amazon has unveiled a new feature in the works for its virtual assistant Alexa that can read aloud into the voice of a deceased loved one based on a short recording of the person.
“While AI can’t take away that pain of loss, it can certainly make their memories last,” Rohit Prasad, senior vice president and chief scientist at Alexa, said at the re:MARS conference on Wednesday. from Amazon in Las Vegas.
In a video released at the event, an Amazon Echo Dot is asked, “Alexa, can Grandma finish reading ‘The Wizard of Oz’ to me?”
“OK,” Alexa’s voice answered.
“Instead of Alexa’s voice reading the book, it’s the voice of the child’s grandmother,” Prasad said. “We had to learn how to produce a high quality voice with less than a minute of recording.”
He added: “We are unquestionably living in the golden age of AI, where our dreams and science fiction are becoming reality.”
Indeed, the feature immediately drew comparisons to fictional depictions of the technology, but ones that are darker than what Prasad was probably referring to, like black mirrorthe dystopian television series which featured an episode in which comparable technology was deployed.
The feature is still in development and Amazon wouldn’t say when it might launch publicly, but its preview comes at a time when cutting-edge AI capabilities are under close scrutiny.
In particular, debate among researchers has intensified over so-called deepfakes, or AI-rendered video or audio to make it look like someone has done or said something that never happened.
It also comes shortly after a Google engineer sparked controversy for arguing that the company’s sophisticated chatbot communicates as if it were sentient, a claim that lacked support from the research community. in AI, but who nevertheless underlined the software’s strangely human communication skills.
Big Tech companies are increasingly studying the impact of AI on society. Microsoft recently announced it was restricting the use of software that mimics a person’s voice, saying the feature could be weaponized by those trying to impersonate speakers as an act of deception.
Subbarao Kambhampati, a computer science professor at Arizona State University, said he hoped Amazon showing a demo of the voice replication tool would make the public aware of the use of synthetic voices in daily life.
“As scary as it sounds, it’s a good reminder that we can’t trust our own ears these days,” Kambhampati said. “But the sooner we get used to this concept, which is still strange to us right now, the better off we’ll be.”
Kambhampati said the Alexa feature has the potential to help a needy family member, though it has to be weighed against a variety of moral issues the technology presents.
“For mourners, it might actually help in the same way that we look back and watch videos of people who have died,” he said. “But it comes with serious ethical issues, like is it okay to do this without the deceased person’s consent?”