I’ve just finished reading this book and thought I’d post a few comments about it. Its written by Blay Whitby, a lecturer in Cognitive Science and Artificial Intelligence from the University of Sussex (ISBN-10: 185168607X).
Overall, I was impressed by this book, as it is very well written, allowing this deep subject to be easily digested. It does not skimp on content, and provides a thorough introduction to the subject at a level that is interesting even for the advanced reader, including discussions on neural networks, evolutionary computation, (un)supervised learning and expert systems.
A key point discussed in the book is the fact that AI is perhaps one of the most misunderstood topics. Most people leap to assumptions – often negative ones where the world is over run with robots!. This book attempts to put these myths into context, and it does this well. In fact, a key point that shines through is that AI is still very much in infancy, so far away from anything approaching that portrayed in films and the like.
It also addresses some deep questions which emphasize the difficulties faced in AI research. Perhaps the most interesting for me is the philosophical question of what intelligence actually is. After all, we have not mastered our own intelligence yet, as we still don’t really know how to measure or determine it (and no, I don’t believe in IQ tests as they are fundamentally limiting a multi-dimensional measurement into a singular value 
). Thus, is the logic that we implement in AI correct or is it fundamentally flawed in some way by our own misunderstandings?
This got me thinking about intelligence, as it can be argued to be a subjective thing, depending on who perceives it. Some people might think that a machine is intelligent if it can fool a human into thinking it is another human, viz. the Turing test! We are getting pretty close to achieving this, yet I don’t think this makes machines intelligent. To coin an analogy, the book describes the so-called Chinese Room thought process, where someone in an empty room is passed cards with instructions on (written in Chinese unbeknown to them) through a small slit from the outside. Based on the symbols, the person inside the box is able to respond to these instructions using their dictionary. To the environment, the box is seen to understand the instructions. In reality, however, this is not the case, as it is simply a matter of syntax without semantics. This is precisely what today’s computers do at their core and this is why there is such a strong argument against AI being truly feesible using current technologies.
Many other questions come to mind in this thought-provoking book, and I would highly recommend it if you have a spare afternoon (only 130 pages).
