Customer Reviews

Understanding Artificial Intelligence (Science Made Accessible)

5 star:		(2)
4 star:		(0)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

This is a collection of Scientific American articles bound together as a book. I read it on the beach. It was great for this and I throughly enjoyed reading it; however, it is by no means an introduction to the field of AI. It's more of sophisticated "gee whiz in the future we have jetpacks" kind of read.

That this book is slightly dated in a very dynamic field seems to matter less than I expected. It covers the basic issues, reviews achievements and goals, and is structured well. As a collection by different authors who may not have written their pieces to be collated into a volume by Scientific American, the book works but is patchy. Thematic elements such as `On Computational Wings' are clear and edifying, whereas more technical topics such as `How Neural Networks Learn From Experience' cover a lot superficially.

I note that Minsky seems to imply that genes do not have repair mechanisms - untrue, but even this great man admits to off days, although SA's editing role should have clarified this point (as a collection of articles I assume it was even twice published). Similarly, sloppy logic/wording when explaining about engineering another hundred human genes to improve longevity detracts slightly from an otherwise delightful article. Minsky's statement "Might not such people, who feel they might not have much to lose (when they die), be dangerous?" is a gem that applies well beyond this topic.

"Will robots inherit the Earth? Yes, but they will be our children." --Marvin L. Minsky

"Understanding Artificial Intelligence" is a great little book (139 pages, paperback). It's one of the wonderful "science made accessible" series by Scientific America's editors.

The book is a compilation of excellent essays by key thinkers in the field. Don't be scared off if you don't have a degree in computer science. All entries are well written and not so heavy as to lose most readers.

This book is a short read but it will stay with you for a long time. If you plan on being alive to see the next few decades, you need to know something about artificial intelligence. This book is a great place to start your education.

The editors write:

"Will a future proclamation be necessary to free artificial intelligence [from human bondage]? Should we allow this to happen? Should we fear out electronic offspring? Will ambulatory AI machines proceed down the Terminator's path . . . or down the benign road become helpful human assistants? Will artificial humans inherit the planet, as some scientists are now inclined to say, or will the melding of biology and bionics simply necessitate a new definition of human? We may not have all the answers yet, but the questions will become more important as each new invention leads toward true artificial awareness."

I highly recommend this book. It educates, yes, but more important, it inspires a sense of wonder and excitement about what tomorrow may bring.


--Guy P. Harrison, author of Race and Reality: What Everyone Should Know About Our Biological Diversity" and "50 Reasons People Give for Believing in a God" (a skeptical analysis of common justifications for religious belief)



The following is a column about robots and AI, published in the Cayman Observer on Sunday in April 2010. If you find it interesting, then you will love "Understanding Artificial Intelligence."


Will robots and AI rule tomorrow?

I am never alone. I share my Grand Cayman home with three humans, two dogs, one cat, and two robots. A few times per week my mechanical friends sweep the floors throughout my house. They navigate around chairs, scurry under tables, and annoy the cat. The more advanced model knows when it's had enough and returns to its docking station to nestle in for a recharge.

This is all very important to me because it means I don't have to sweep. While the robots work, I am free to engage in activities far more appropriate for a member of the most intelligent species on Earth--like watching TV.

Robots have plans that extend far beyond mere housework, however. Soon they will be everywhere. But there is the slight possibility of a downside--extinction of the human species. Yes, intelligent machines in the not-so-distant future may become so smart that they could threaten our existence. My cute little floor sweepers may turn out to be the great-great grandparents of real-life "terminators" that will kill us all one day.

Just like any other normal Cayman family, my children and I often discuss the impending robot apocalypse. My son reassures me that robots won't have any need to fight us because we will eventually be machines just like them. We will become robots, he believes. Given recent advances in prosthetic limbs, ear implants, brain implants, and so on, he may be on to something. Maybe we will merge with our technology so intimately and thoroughly that there will be no "us" and "them" to define battle lines. My daughter is not worried either. She is convinced that someone will be smart enough to remember to simply program the robots to be nice. Ah, don't you just love youthful optimism?

Still, I wonder. With computers on course to become freakishly powerful in about three or four decades, and with robotics development in high gear, will we be able to hang on to civilization's top rung? It seems likely that a game-changing new "species" is on the horizon and approaching fast, one that will be difficult if not impossible for us to control. Thanks in large part to unprecedented military investment in robotics, we are now stepping into a very different world--for better or worse. "In the blink of an eye," writes Peter Singer in his book, "Wired for War", "things that were just fodder for science fiction are creeping, crawling, flying, swimming and shooting on today's battlefields. And these machines are just the first generation of these new technologies, some of which may already be antiquated as you read these lines."

Hugo de Garis, an artificial intelligence researcher as well as my pleasant Face Book friend, may be engineering our collective doom. He is the author of the nightmare-inducing book, "The Artilect War," in which he admits feeling conflicted about his work. He believes it is highly probable that super-intelligent machines will brush us aside one day. Despite those fears, however, his research is so fascinating that he can't stop himself.

Unprecedented transformations are occurring right now. For example, did you know that in 2009 the United States Air Force trained more ground-based "pilots" to fly robot planes than it did traditional pilots for conventional planes? This year, the Air Force is projected to acquire more new robot planes than new conventional aircraft. This represents a monumental shift in the human-robot equation, yet the public mostly doesn't know or doesn't care. The various military robots that we know of are controlled and monitored closely by humans today, but what about tomorrow? Robot autonomy on the battlefield will be here soon, if it isn't already.

The military, rather than my floor-sweeping needs, is driving much of the cutting edge research. So much so that when the final histories of the wars in Iraq and Afghanistan are written one day it is possible that surprisingly little attention will be given to Bush, Saddam, the Taliban, oil, and terrorism. The greatest impact of these wars may well turn out to be robots.

I attended an exhibition of advanced and near-future technology called "NextFest" a few years ago in Chicago. I spoke with a representative from the company that makes one of the drones used by the US Air Force. I met a solider who wore a sleek prototype suit that, when fully developed, would make him stronger and even tighten automatically like a tourniquet to slow blood loss if he was wounded in battle. A Japanese woman demonstrated how strong she was thanks to the robotic "exoskeleton" she wore. I saw ASIMO, the famous Japanese robot, do its usual slow shuffling walk. Interested but not overwhelmed, I felt like I was watching the Australopithecus of robot evolution. But, of course, the robot equivalents of Homo erectus and Neanderthal are already being designed or built somewhere right now. There will be no four-million-year-wait for them. Don't blink.

I am not suggesting that anyone should panic or lose sleep over a possible robot takeover in the future. At the very least, however, you should be aware of what they are up to. The robots are not coming; they are here already. The invasion has begun. I know because an indifferent little robot on a mission just rolled by in front of me right here in my living room. It cares nothing about me; it just wants to sweep. That's what it does; that's who it is. I admire its dedication and focus. But I wonder, will the day come when its descendants demand more?