Drew HarwellJan 10, 2023

This will only be a problem if someone gets a three-second recording of your voice

https://arstechnica.com/information-technology/2023/01/microsofts-new-ai-can-simulate-anyones-voice-with-3-seconds-of-audio/

Microsoft’s new AI can simulate anyone’s voice with 3 seconds of audio

Text-to-speech model can preserve speaker's emotional tone and acoustic environment.

Ars Technica
Show threadskry

@drewharwell "it is possible to build a detection model to discriminate whether an audio clip was synthesized by VALL-E."

I wish they had said clearly that they were going to build that and require it stay detectable. All the banks wish that too.

Jan 11, 2023 at 6:07amWeb