Comment by ahartmetz

They could at least try to vaguely match the voice and maybe cadence of the original. AFAIU it's one of these things that would have been too hard ten years ago but is fairly easy now. Too computationally expensive probably.