The way I see it, as a reader mind you, is that so long as the person's name is out there I can decide for myself whether or not to consider the person an expert.
Well, yes, that works as far as determining what counts as an expert. It doesn't work as far as verifying someone's qualifications. We can't expect readers to phone the university and confirm someone is really a lecturer there for each article they read, whereas we could do that ourselves if we deem it necessary. Also, there are ways we can confirm the person we're talking to really is Professor Joe Bloggs (email from official address, putting a note of official website, etc), a reader can't do that.
But yeah, I guess I misread your suggestion. "What experts should do as experts is review articles and put their stamp on them as being correct." I guess the problem with that is that the articles have to be correct first.
You're right, some system for experts to correct or suggest corrections to an article would also be required.