I've noticed questionable review data appearing in the SERPs of late too, and I'm frankly rather surprised. You can see an example [plus.google.com] on a Google+ post I wrote a short time ago (and I've since seen similar types of reviews generate rich snippets).
I'm surprised because Google has traditionally been less rather than more trustful, yet - at least when it comes to producing rich snippets - it seems to be processing review content almost on face value.
I trace this back, at least in part, to January 2012, where a post [insidesearch.blogspot.ca] on Inside Search listed this as one of their "30 search quality highlights":
More rich snippets. We improved our process for detecting sites that qualify for shopping, recipe and review rich snippets. As a result, you should start seeing more sites with rich snippets in search results.
The interesting point here revolves around the declaration that this has to do with "detecting sites" - that is, review data may not be evaluated on the individual merit of a review's trustworthiness, but on the overall trust in the site. So "trusted" sites are capable of marking up review content with schema.org microdata and being granted rich snippets for their efforts, without other algorithmic oversight.
What I've found especially surprising is that Google seems to now being routinely honoring review markup, even when the provenance of those reviews is problematic. That is, rich snippets are being generated even in situations where there's no discernible path for a user to even submit a review.
I think this is all related to Google's efforts to promote structured markup, and particularly schema.org microdata (which is what I've almost exclusively seen in the code of sites with "questionable" review rich snippets). A new technology obviously sees the best adoption when there's a demonstrable benefit to using it (witness the rapid ubiquity of "Like" buttons and underlying OpenGraph data, because it obviously improved brands' visibility on Facebook), so Google has been keen to return rich snippets as a incentive for webmasters to markup their own reviews.
But I think it's clear that Google has been too generous in according sites "trusted for reviews" status, and I think we can expect to see that trust level dialed back - or additional algorithmic methods put in place to accord for the quality of review data itself, rather than only assessing a site's trustworthiness.