Using AI to recreate a real person’s voice is the dumbest possible use. It’s like drawing a cartoon that can only resemble a real living actor. Just… make something up. How does this character sound, in your head? Generate that, and then use style transfer, so anyone can do that voice.

Because the text-to-speech version is only the voice… not the character. You still want real human actors performing the character. The AI part is just a costume for their throat.