Haber says that while his system allows real-time, speech-synchronised facial animations, this is probably not useful for forensic identification because it might confuse people. "You don't want to show talking heads," he says.
Well, perhaps not for forensics, but I can think of lots of other fun applications. Good speech synthesis, a conversational AI, and this, and we have Talking Deads. Who wants to bet someone couldn't make millions with the idea?
artistic