Even if the transcription was instantaneous, you'd still need to manually vet it, i.e. listen to all of it.