Here is an example of a question with an accepted answer that should not have actually been accepted:
Granted, the question itself isn't very good because it's asking about multiple referred claims. The comments are essentially correct in that the subject is not on a crusade against high-fructose corn syrup, but sugar in general. Having read his book, if I recall correctly, he doesn't really make a differentiation between HFCS and other forms of sugar.
It doesn't seem to me like this answer should have ever been accepted in the first place. Has there ever been a case where a moderator has unaccepted someone's answer, or is this something we will ever do?