Zero shot implies that it was given no direct examples [0] So, in this case, it wasn't given any examples of the exact voice in combination with text, it is just using the prompt voice + a prompt text to generate new audio.
To me, it seems that the only AI generated content that isn't zero shot would be the narrow subset of generations where it has multiple training examples for the exact requested prompt. i.e. anything that is composing multiple pieces of information together is "zero shot".
How should I parse 'shot' in the context of this term? Shot makes me think 'attempts', seems like a weird word to use for ~= "training examples"
one shot - given the text "run faster" along with Alan Greenspan's voice pronouncing that phrase, the model can produce Alan Greenspan's voice saying any other phrase
zero shot - given only Alan Greenspan's voice pronouncing "run faster" but no text version of what was said, the model can produce Alan Greenspan's voice saying any other phrase