I think it's hard for us geeks to understand why art is valued the way it is. If we have a wicked computer that works perfectly, but later find out that it's (omg) not a real AMD, but in fact some sort of replica... we say "who cares?... As long as the IO behaviour is what I wanted, that's all that matters!"
Seriously, though... like it or not, the way that art appraisal work has little (nothing?) to do with talent of the artist, and everything to do with perceived value and context. After all, a sufficiently awesome printer could produce (in principle) copies of Rembrandts that most people would not be able to differentiate, but ultimately they would have little value. Similarly, artificial diamonds are just as good (or better) in terms of purity, hardness and optical properties as natural diamonds, but the natural diamonds are valued higher "just because."
Even if no one can tell that it's a fake for many years, art critics want to know if it's real or not. Such knowledge can change the perceived value of the item, even if it doesn't change its physical appearance. Again, art value is NOT about how "nice" or "well done" a work is, but rather based on "how much are people willing to pay for it."
And in a strange way, having some Rembrandts shown to be fake would actually INCREASE the value of all the other Rembrandts, since they would suddenly be perceived to be a more rare commodity than before. So in fact a Rembrandt collection could stand to have its calculated worth INCREASED if some of them were found to be fakes. (Obviously other Rembrandt collections would also increase in value, especially if it were found that they contained no fakes.)
Lastly, let me mention that above and beyond the determination of the value of art, it's worthwhile from the perspective of art history to determine which ones are real and which are not. If a given conclusion about a time period is based upon a painting that turns out to be fake, well then we have to update the (art) history books.