rolph 14 hours ago

cloning the voice is only step one, easy part.

peculiarities of syntax, and grammar must be emulated, and of course any safewords, or authentication phrase must be elucidated, you need more than 3 seconds of speech to do this.