With humans (and some machines) we tend to put/manadate additional processes for certain risks instead of just relying on their own good nature. Why just rely on the machine when elsewhere we have learned not to necessarily just trust them so much?