The Technological Singularity

Murray Shanahan

Published by
The MIT Press
ISBN 978-0-262-52780-4
RRP £11.95
Reviewed by Danny Williams MBCS CITP
Score 7 out of 10

Artificial Intelligence is not new - it is something that people have been working on for decades. It has been the topic of many books and films and an area that scientists, and technologists have devoted their careers to. The singularity is a time (still in the future) when the capability of artificial intelligence surpasses human intelligence and become self-sustaining.

In this book the author discusses how the singularity might come to pass and what the ramifications might be. I would characterise the book as a fusion of technology, philosophy and science fiction. The fiction part might seem a little unfair but as the book discusses things that have not yet happened it's hard to know whether they actually will or not.

A decent amount of the book is dedicated to explaining how different types of artificial intelligence could be created. One suggested approach is to replace each neuron in a human brain one by one with a technological equivalent. Eventually you have an entirely artificial brain that theoretically would be indistinguishable from a biological one.

I enjoyed how the author builds his arguments layer by layer, where each step seems to be plausible and builds on the previous layers.

The concept of a virtual society populated by artificial intelligences was quite intriguing, as was the idea of artificial intelligences running projects to develop artificial intelligences! The author discusses how we should treat artificial intelligences. Should they have the same or even better rights than humans given what they are capable of? Should they be allowed to vote? What if an artificial intelligence created a large number of other artificial intelligences just for the purposes of swinging an election and then destroyed them?

Some humans use chemicals to enhance their performance - from caffeine to illegal drugs. The author suggested that artificial intelligences could do the same thing - effectively ingesting artificial performance enhancing drugs. I presume there would be gangs of artificial intelligences designing, manufacturing, distributing and selling these drugs too! You can see how some of the ideas can end up in the more fanciful end of the reality spectrum. But the journey never hurts your brain (real or artificial!)

If you'd like to suspend reality for a while and think about what the future of artificial intelligence might be like then you'll enjoy reading this book. Just don't make any plans based on what you learn.

Further information: The MIT Press

June 2016