You may also be interested to learn about the various review queues and the like. Right now there are some limited sanity checks (post length, certain words triggering a weight filter, etc) that suggest if a post may be good or not, but then we're also training the system "Mechanical Turk" style by having humans go through and audit thousands of posts. Eventually you would be able to apply some ?bayesian learning? against the thousands of posts that have been previously manually filtered and use those as a trainer for the "bot" to catch some of this stuff.
However, the actual situation on all Stack Exchange sites is that this is about the human exploration of ideas. You want to learn, to interact, and to grow. Having an automated process do anything more than "your post is too short" or "here are 15 duplicate questions covering the same ground as yours, do they help you or is there something unique about your situation that hasn't been covered yet?" ... is really going to be anti-Stack Exchange. We aren't a Mechanical Turk Encyclopedia, even though that is a portion of the point of the network. What we really are, and what we really want to focus on, is a lot of humans interacting like humans.