False pattern recognition is a factor in racism. That is someone sees an obvious physical characteristic and associates it with a negative event. Because the human mind is always looking for patterns it sometimes picks out ones that are not really there. What I wonder is if this will also be true of robots in some fashion. A robot might have the same correlation does not equal causation problem and so might, falsely, pick out some physical characteristic to decide things on and then run amuck with it. Perhaps it would pick out glasses wearers or non-wearers as something and then cause a confirmation error by proceeding to treat them all with that assumption.