Speech Recognition is something I would be more concerned about. Amazon and Google with a decidedly more “liberal” stance on data collection have more raw material to train and validate models on.