I haven't read the paper but if Wikipedia serves as a detection tool, I don't particularly see a problem using it. Perhaps the real problem is the non-Wikipedia, canonical systems to find these probably-relevant cases?