The Truth According To Wikipedia
I think that as time goes on, the scope of our knowledge will be split into three parts.
First, pure mathematical knowledge, for instance the proof that there is an infinite number of prime numbers. A knowledge which is independent to human motivation or primary perception. Usually that knowledge is eternal. In this first part I would also be tempted to put things like evolution. I know that there is quite a hot debate nowadays about evolution, but this debate is irrelevant. Either evolution did happen or it didn't. If it did happen we will have the proof one day (possibly by inventing time travel machines), and if it didn't we will also know it. I would not put in this first category the question of deciding whether global warming is going to happen (I mean more than it s already happening). In this last case, it is not a matter of just sit down and wait and see whether it will happen or not, because our actions, what we do now (or not), can determine whether it happens or not.
Second part is our scientific knowledge, such that the correctness of Newtonian mechanics at non relativistic speeds. A knowledge consistent with our experience and perception, but which, given the evolution of our capabilities of perception, may under go re-factory and reformulation.
Third part is the rest. The 6 and half billions different versions of truth that we can find on the planet.
So we should not debate whether or not wikipedia is a good idea. What we should is have three of them. The first one for the first kind of knowledge, which can be edited mainly by people with the relevant education (I do not see how someone without any mathematical education could edit the entry on the Fermat's last theorem). I also guess that this first encyclopedia will have additions, not corrections (given that if something was incorrect it would not have been there in the first place). A second encyclopedia for the mainstream scientific knowledge, which should be edited only by university experts. And then the third one, the one for the rest of human knowledge (anything not mathematical, not checkable and not scientific), editable by everyone at will. The one that you cannot trust, but which is better than any other form of encyclopedia that we have come up with anyway.
... and then one day, when Google and the AI folks will have done enough work, dedicated artificial intelligences will take over the writing of encyclopedias. And this all encyclopedic mess will be over.