Three seconds. That's all a voice cloning AI needs to sound like you.

Not your life's work. Not hours of training material. Three seconds of audio — a short greeting, a voicemail message, even a video of you speaking — and software can now create a convincing replica of your voice saying anything. Anything at all.