I agree with this to a certain extent. Yes, it may well not be reasonable for an examiner to make a rejection based solely upon material published on Wikipedia, especially when there are no related and/or supporting references cited. In such cases, I think that an applicant would be fully justified in complaining. I personally would use Wikipedia entries as supporting references, but not directly for rejections, if I could help it (and I've never needed to yet). I find that Wikipedia is useful in that there are often links to other relevant documents, documents whose credibility is not in question.
I don't agree, however, with those who would forbid the use of Wikipedia for any purposes. It makes sense to me to allow the use of Wikipedia, but to give it the weight it's due. Prior art is nothing more than evidence, and like any other evidence, it needs to be evaluated based upon its credibility. Everyone knows that the credibility of the authors of Wikipedia entries is generally unknown and uneven. But it does contain quite a bit of good information, and ruling out its use as prior art out of hand seems to me to be overkill.
I also have seen some talk about the fact that the Internet Archive, better known as the Wayback Machine, contains a disclaimer regarding the accuracy of the contents of its database:
"You understand and agree that the Archive makes no warranty or representation regarding the accuracy, currency, completeness, reliability, or usefulness of the content in the Collections, that the Site or the Collections will meet your requirements, that access to the Collections will be uninterrupted, timely, secure, or error free, or that defects, if any, will be corrected. We make no warranty of any kind, either express or implied."
Once again, the discussion was about whether the Wayback Machine can be relied upon for prior art to be used by examiners in rejecting the claims of a patent application. And once again, I'd say that you need to consider the material for what it's worth. In the case of the Wayback Machine, however, I think I'm pretty safe in saying that the material is somewhat more reliable/credible than that from the average Wikipedia entry.
Since the Wayback Machine gathers its data through the use of automated robots, the only way (that I know of) for the data to be unreliable is either that their software is buggy (which doesn't strike me as extremely likely) or their database is getting hacked (which also strikes me as unlikely). If the web content in the Wayback Machine's database gets corrupted, then it doesn't seem to me as if it would be useful as prior art, since it likely wouldn't even load and display. If the database is getting hacked, then you've got to ask yourself why. Better yet, it seems extremely unlikely that someone would rewrite old web pages in such a way that would make it useful to an examiner, unless you can convince me that some competitor of the assignee of the application under examination is doing it. Talk about unlikely.
I just can't see a simple disclaimer as grounds for excluding the contents of the Wayback Machine as a source of prior art for examiners. If it ever does happen, though, then you can look forward to more pretty obvious Internet patents to issue, what with the speed at which Internet companies (and their web pages) evolve and/or go out of business. Without the Wayback Machine, examiners would have a much harder time trying to establish what actually existed on the Internet in the past.