Alexa Vice President and Chief Scientist, Rohit Prasad, presented his vision for the future of the well-known voice assistant and that vision includes the fact that the machine is capable of emulate the voice of its owners. As well as to resurrect the vocal timbre of loved ones by now deceased.
Like a modern necromancer, Alexa summons the dead
This bizarre – and not a little controversial – approach to technology was touted with great fanfare at Amazon’s annual Re: Mars conference, which is currently being held in Las Vegas. Among the new features presented for the occasion, the one that announces the possibility for the tool of synthesize the voice of humans to replicate it when most appropriate.
If that weren’t already disturbing in itself – if only because it recalls the death traps seen in the Predator film series -, the company has decided to promote this unprecedented Alexa talent explicitly alongside the dead, or representing him in the role of a departed grandmother who reads a story to her granddaughter. “While AI can’t eliminate the pain of loss, it can definitely make sure memories survive over time,” Prasad said.
The technical details regarding the function are currently contained, however we know that the speech synthesis procedure is subject to extremely limited training. A one-minute recording should be sufficient to replicate the mimesis subject with some fidelity. At the same time, it is easy to believe that, in an attempt to avoid increasingly current controversies, Amazon will decide to limit the function to the mere sound level and therefore does not want to also simulate the spirit and character of the family members who have passed away.