Great point -- it will be interesting to see how one becomes a "trusted" reviewer -- mainly, whether requires knighting by an admin or an algorithmic judgment a la auto-confirmation.

In the article I wrote about Wikipedia last year, I mentioned the interesting model at Helium for refereeing content. It wouldn't work for Wikipedia, but I believe they do measure authority there by subject matter -- i.e., you're invited to referee submissions on subjects similar to what you've written about previously.