More broadly, audio tends to be the human-preferred input for AIs while vision (images/animations/video) is the preferred output.