In argument people often say "you are generalizing". The implication being that you should never generalize. This is the equivalent of saying that drawing statistical correlations are not useful, and is clearly nonsense. However if people object to overgeneralizing then they have a point. "Most loud cars are big and you have loud car, therefore it must be big" is nonsense. The predictor is unlikely to be absolute.

Few people make this distinction. Particularly those who are trying to selectively discriminate.