But does it also do speech recognition (not just command recognition) and voice synthesis? It would break the immersion if everyone in a space station bar would look like aliens but sound like humans. Also of course playing a character of opposite gender is of interest I'm sure. (The male actor in Mass Effect was so dreadful I played both games through with the greatly voice acted female Shepard)
TTS and STT is not new but realistic, emotional expressive speech (or singing) synthesis (and analysis) is not exactly where we seem to be today or in the next couple of years for that matter. It'll be very cool when we get there.