Super-Intelligent Machines (Ifsr International Series on Systems Science and Engineering) [Hardcover]

Bill Hibbard (Author)

Book Description

ISBN-10: 0306473887 | ISBN-13: 978-0306473883 | Publication Date: October 31, 2002 | Edition: 1

Super-Intelligent Machines combines neuroscience and computer science to analyze future intelligent machines. It describes how they will mimic the learning structures of human brains to serve billions of people via the network, and the superior level of consciousness this will give them. Whereas human learning is reinforced by self-interests, this book describes the selfless and compassionate values that must drive machine learning in order to protect human society. Technology will change life much more in the twenty-first century than it has in the twentieth, and Super-Intelligent Machines explains how that can be an advantage.

Product Details

• Hardcover: 236 pages
• Publisher: Springer; 1 edition (October 31, 2002)
• Language: English
• ISBN-10: 0306473887
• ISBN-13: 978-0306473883
• Product Dimensions: 8.9 x 6.2 x 0.8 inches
• Shipping Weight: 1 pounds (View shipping rates and policies)
• Average Customer Review: 4.0 out of 5 stars  See all reviews (1 customer review)
• Amazon Best Sellers Rank: #5,338,528 in Books (See Top 100 in Books)

More About the Author

Discover books, learn about writers, read author blogs, and more.

Customer Reviews

Most Helpful Customer Reviews

1 of 1 people found the following review helpful:

4.0 out of 5 stars A decent primer on the future of super-intelligent machines, June 29, 2007

By 

Will Tanizaki (New York, NY) - See all my reviews

This review is from: Super-Intelligent Machines (Ifsr International Series on Systems Science and Engineering) (Hardcover)

Hibbard says we are entering another kind of Gotterdammerung--a birth of new gods. This is just another way of saying we will eventually build super-intelligent machines, ones that far surpass the intellectual capabilities of human beings in every sphere.

This book is well worth reading. However, I wouldn't put it in the same category as Ray Kurzweil's "The Age of Spiritual Machines". In terms of similar books, I think J. Storrs Hall's "Beyond A.I." has more to offer, as does Joel Garreau's book "Radical Evolution" (not specifically about super-intelligent machines). That said, I wish there were more books like this one on this topic. I'm glad I read "Super-intelligent Machines". In general, I trust Hibbard's basic exposition of the technological factors involved, but I found the book somewhat unsatisfying, mostly because I think Hibbard oversimplifies too many complex issues. Hibbard, a researcher in artificial intelligence, offers a clear sketch of coming technology, but he stumbles when discussing the philosophical issues involved in creating super-intelligent machines.

Those familiar with the idea of super-intelligent machines will see the basic territory covered, as Hibbard retreads much ground covered by Kurzweil and Hans Moravec. The book is as good a place to start as any for the person who hasn't even considered the possibility that there could be super-intelligent machines in the 21st Century. More than simply a rehashing of Kurzweil, "Super-intelligent Machines" offers some fresh perspectives on what the future might be like when this threshold is passed. If anything, Hibbard is too conservative about the implications of greater-than-human intelligence. One of the problems with writing a futurist book like this is that it may be impossible, in our current state, to comprehend what life will be like on the other side of that divide. So some of Hibbard's musings about future societies seem parochial. What would be useful would be some sustained argument and some creative extrapolations concerning the direction of our human-machine civilization in the next few decades. Too often, Hibbard makes declarations about the way the future will play out, while failing to take into account potential counter-arguments.

I was also hoping to see some thoughtful discussion of the philosophical issues involved in machine intelligence, something beyond Searle's Chinese Room thought experiment. Here I was disappointed, particularly by Hibbard's treatment of machine consciousness. In treating the question of whether super-intelligent machines will be conscious Hibbard writes: "The problem of consciousness has been fundamental in the development of human philosophy. Plato speculated that the external world is just an illusion, and that only our internal mental world is real. The dualism of Rene Descartes asserts that the gap between subjective and objective cannot be bridged. Understanding human consciousness in terms of the physical behavior of neurons will bridge this gap." I happen to believe something along the lines of what I think Hibbard is proposing (I think this bridging may come about as a consequence of super-intelligence); but his handling of this issue is less than deft. First of all, Plato did not hold that "only our internal mental world is real"--this would be a kind of idealism. And, although Descartes's interactionist dualist "solution" may be unsatisfactory, he averred that the subjective and the objective could be bridged by means of the pineal gland. The idea of an "explanatory gap", while indebted to Descartes, arises from 20th Century philosophy of mind; it is often credited to Joseph Levine. I can't help but feel there are other parts of the book that are also put forth in a somewhat unsophisticated fashion. These are minor faults, but they detract from the overall value of the enterprise.

Hibbard does make some important points. He points out that "the education of super-intelligent machines should be monitored by impartial (i.e. independent of the organizations building the machines) public representatives to make sure the machines are emotionally healthy and happy, with a good attitude toward serving people." (p.142). Although I think the use of phrases like "emotionally healthy and happy" smacks of anthropocentrism, the general point is right on. It is not OK for the first company/ government agency that designs a super-intelligent machine to police itself. It's too bad that there seem to be so few people who really grasp all the general technical issues involved (and I am not one of them). Hibbard is clearly someone familiar with the design of artificial intelligence, so his perspective is welcome. One omission is that Hibbard doesn't delve into the questions involved in developing so-called "friendly AI", or artificial intelligence which acts in accord with our highest ethical principles. But his point is well-taken: it doesn't make sense that we have hundreds of people overseeing the introduction of a blockbuster new drug, while it is scarily likely no one will oversee the creation of the first super-intelligent machine.

Hibbard is optimistic about the unfolding of super-intelligence. As he sees it, the benefits of building super-intelligent machines will be too attractive for humans to resist. He sees super-intelligent machines largely replacing the work of (unaugmented) human beings. This massive unemployment will lead governments to support the needs of everyone, perhaps by extending social security benefits to citizens as soon as they are born, as has been suggested by Hans Moravec.

Unfortunately, any book in this genre will be compared to the work of Ray Kurzweil, and in this comparison, I'm afraid Hibbard comes out behind. Kurzweil does a good job of laying out the big picture in "The Age of Spiritual Machines" and of giving a survey of the details in "The Singularity is Near". Hibbard seems too often to settle for pronouncements about the way that things will turn out, and these pronouncements don't seem to have the requisite arguments supporting them. Nonetheless, the book is a quick and interesting read.