Speech to text software alone isn't good enough, it makes too many errors, the DoJ discusses the evidence of that. Captioning video is basically a solved problem, through outsourcing.