Book Review

Book Review: The Cambridge Quintet by John L Casti

There have been some exciting books published recently in the field that could be known as artificial intelligence. The field and what they've tried to accomplish is a moving target, and with some of the recent advances and publications I've decided to check back in. What started out as a question for computer science has been passed mostly to neuroscience, and is more of an interdisciplinary research area along with physics and robotics, as well as others. In preparation for a little reading tour this year of this fascinating subject, I picked up again “The Cambridge Quintet” by John L. Casti.

Sometimes, the best place to start is at the beginning.

The premise of the book is a dinner party that never actually happened, but you wish did and that you were invited. The setting is post-world war II England, and people are just beginning to understand what computers are and are capable of doing, having been instrumental in code breaking during the war.  Alan Turing has just floated the idea that a computer can be built to mimic a human mind- a fantastic notion. The author ­C.P. S­now hosts a dinner party to thump this idea. He invites a cadre of leading linguistic and scientific experts and philosophers to debate Turing's idea. The author sets his real life characters (the dinner participants are all real people from various disciplines in the early 20^th century) into this fictional setting and lets them have at it. The description of the food (each chapter is a course) is an added bonus.

Alan Turing is known as the father of artificial intelligence and is most popularly known for his idea of a Turning test, which is what the dinner guests thump. The idea behind a turning test is roughly as follows: that if you were conversing with someone or something behind a wall, and the answers to the questions could convince you that the other person was a live, sentient human but it was really a computer program, then you could claim that the machine was a working replica of a human mind.

This claim of course, sets off a plethora of debate of what it means to be human, have language, have emotions, etc. etc. which was all fascinating and a good refresher for the start of my reading tour. It’s a great introduction to this field, and Casti does a good job of having his dinner guests debate their view points. Casti is well enough versed with everyone’s arguments to come up with some novel conversation points. He has Turing concede that if ESP was invented, that his test would be invalid. I hadn’t heard of that before, and it was both novel and made me laugh.

At the time, the idea of a machine passing the Turing’s test was tantalizing and so investment into the field was made, of which we reap the benefit today (not least of which I can bore you with my blog and book reviews). Massive leaps in computing power along with networking have enabled ever complicated and far reaching applications. But as for Turing’s test - well the story here is more complicated.

The one outcome that the dinner guests didn’t count on was this: It doesn’t matter. If you were able to have a computer do polite conversation for 10-15 minutes – so what? You haven’t learned anything about the human mind. You’ve done the equivalent of teaching a horse to count. The things computers are good for – boolean logic tasks like routing traffic or calculating the odds at chess are impressive, but not what makes someone human. To replicate how a human mind works, you have to know how a human mind works. So about 10-15 years ago or so it was like all the computer science artificial intelligence people said collectively, rather roughly, “okay we are going to hold that thought while we go figure out neuroscience and neurobiology.”

So brain research is the #1 area of NIH funding now – not only to answer these questions but also because we don’t have effective treatments for brain diseases – Alzheimer’s being first and foremost on the list.

But there is still other work to be done while the brain research people do their thing. For example, the robotics people have their work cut out for them. Turing had his fictional computer behind a wall, but an important aspect of being human is interacting with the environment. This is a point the philosophers at Casti’s fictional dinner bring up. Replicating the senses and processing them and writing programs to appropriately respond to them are a herculean scientific challenge.

A lot has changed since the end of World War II and with any nascent field there is bound to be a lot of self- questioning, back peddling and good old fashioned angst. Nevertheless the optimism and can-do spirit of those in the field is one of the reasons I’m going to try and check back in this year (though I can’t promise not to get side tracked) – and a spirit Casti captures in his book.