Amazon’s Alexa voice assistant has been able to tell children bedtime stories for quite some time now, but the company’s latest innovations could help to bring beloved storytellers back from the dead.
In a demonstration at Amazon’s re:Mars conference, the company was able to show how a brief recording of a deceased loved one’s voice could be synthesised to read the entire story from a device like an Echo speaker.
The demonstration (reported by TechCrunch) required just a one minute recording, with the output described as “very impressive” by reporters in attendance.
“This required inventions where we had to learn to produce a high-quality voice with less than a minute of recording versus hours of recording the studio,” said Amazon’s Senior Vice President and Head Scientist for Alexa, Rohit Prasa.
“The way we made it happen is by framing the problem as a voice conversion task and not a speech generation path. We are on questionably, living in the golden era of AI, where our dreams and science.”
How far this tech is away from reality remains to be seen, as details weren’t forthcoming during the keynote stream. However, as magical as this sounds, it’s likely to raise all manner of ethical issues about the inability for participants to consent to the use of their voice, synthesised or not. Needless to say it would also bring joy to millions who miss the sounds of their loved one’s voices.
The re:Mars conference (which stands for machine learning, automation, robotics and space) features a number of keynote addresses from top speakers in those sectors. You can watch some of the keynote events here at the re:Mars website.