You could look at how AI has transformed over the past 2-3 decades. The focus has shifted from developing general (i.e human like) intelligence to using statistical method to solve practical problems.
Keywords you may want to look into: machine learning, neural networks and probabilistic reasoning.
Take a look at some papers by Sebastian Thrun or Andrew Ng....
http://scholar.google.com/scholar?q=...r=&btnG=Search
http://scholar.google.com/scholar?hl...un&btnG=Search
Specifically you may want to read Andrew Ng's paper on robotic helicopters, and maybe the darpa challenge paper by Thrun.
A book I would suggest is "Artificial Intelligence : A modern approach"
As far as classifying the "true nature" of AI ... I think this is difficult to do, since there is really no exact definition for what "intelligence" really is. AI is currently viewed in CS as the ability to solve complex problems that generally require human "reasoning" ... the methodology used by the computer to do this is irrelevant... most of the time it is just a really good statistical/guided guess after looking at many possibilities.