I read this book only recently, having read the author's two most recent books "The Age of Spiritual Machines" and his book "Kurzweil vs The Critics of Strong AI". Both are excellent books, and reflect the author's extreme optimism about the future of artificial intelligence. He is definitely one of the best apologists for AI, and documents well its living history. Reading this book after the recent ones gives an interesting comparison between what was real in AI then and what is real now. Indeed, the AI landscape has changed dramatically, and there were a few companies specializing in AI in business at the time of publication of this book, that are not around any longer. But for every company that has failed, there have been many more to take their place. Their character as companies has changed, due in part to the rise of the Internet. In fact, it is network engineering that has resulted in many of the applications of AI in the last 5 years, and those applications of course are not mentioned in this book, due to its date of publication.
The author begins the book with a discussion of what he calls "The Second Industrial Revolution", which, he claims, is now in progress, and is based on the rise of thinking machines. These machines will extend and leverage human mental abilities, he says, challenging the human uniqueness in this regard. He expresses caution over the idea of making our military defenses controlled by intelligent machines, at the same time expressing his confidence that machine intelligence will indeed be sophisticated enough for this to happen. This revolution is here he says, will be more radical than the first one, but cannot be stopped, and he encourages therefore the constructive use of its technology. Thus is the author's motivation to write this book: to give the reader an overview of what was possible in AI at the time, and encourage the benevolent use of it.
The author not only discusses the technology of AI, but also attempts to give the reader insight into just what AI is. This entails a discussion of philosophy, since philosophical debate dominated AI in its early years. Such debates are still common, but due to the frequent vituperation involved in them (which the author recognizes and mentions in the book), not much is to be gained from these. Time is better spent on actually trying to build thinking machines, and not engaging in conversations that lead nowhere. Since this book appeared, many philosophers have left their "arm chairs" and have joined in the practical research in artificial intelligence. This trend will no doubt continue in this century, thus giving rise to the "industrial philosopher".
A fairly detailed history of the field of artificial intelligence is given in the book, with several articles written by some of the more recognized individuals in the field. All of these are interesting reading, and shed light on the different attitudes and prejudices regarding AI. For readers who are new to AI, this will be welcomed, as well as the many discussions on the mathematical foundations of AI and its intersection with cognitive science.
The author refrains from including any mathematical notation or equations in the book, and this has its advantages and disadvantages. It allows a more general readership but sacrifices some of the clarity of thought that mathematics allows. The author does give a good discussion of pattern recognition though, especially edge detection. His discussion on this topic is interesting in that it brings up his demarcation between "logical" and "parallel" thinking. Logical thinking is referred to as "sequential" and "conscious", with a resulting limitation in computational ability. It is to be distinguished from parallel thinking which can process multiple levels of abstraction, and can occur without conscious direction. Pattern recognition is in his view an example of the latter, and he justifies this view in the book in some detail. More evidence for his view from laboratory experiments is needed however. Pattern recognition algorithms and technologies have exhibited considerable advance since this book was published.
There have been many advances in AI since the time of publication, due in large measure to the rise of the Internet. Most of these advances have been breathtaking, such as in computer chess, games with imperfect information, Bayesian networks, financial engineering, network intelligence, literary creativity, automatic theorem proving, to name just a few. The author discusses his projections for the future of AI in the book, and it is interesting to compare them with what really came about within the decade later. There is no doubt that more exciting developments are on the way, and the optimism expressed by the author in all of his writings is also characteristic of all who are responsible for these developments. The machines, getting more intelligent with every decade that passes in the 21st century, will bear the signature of these individuals: a tell-tale sign and proof of the genius of the human species.
