What Arlindo Oliveira sets out to do is certainly broad in sweep. He gives us background chapters on the development of electronics, computing, AI, cells, the brain and more, then brings them all together in a synthesis that examines the possibilities and implications of artificial minds, whether limited - for example, does Google have a kind of mind? - to being fully conscious. Without doubt there's a lot to interest the reader here, particularly once Oliveira gets to the synthesis part.
Of the introductory bits, not entirely surprising given Oliveira is a computer science professor, the computing parts probably work best. The biological parts seemed rather dull to read, and though there's plenty of material there, it certainly wasn't the best introduction to cells or the workings of the brain. However, the reader who persists will be rewarded with genuinely interesting material on how we should treat an artificial intelligence, what the implications of copying a digital intelligence are and so forth. Interestingly Oliveira did not regard the concept of a conscious AI as 'speculation' - he left that to the Singularity.
Perhaps the most worrying part was some not entirely accurate history of science. We are told 'Later in the nineteenth century, punched cards would be used in the first working mechanical computer, developed by Charles Babbage' - but unfortunately, they weren't, it was never built. We are also told a working version of Babbage's Analytical Engine was made in 1992 and is on display in the Science Museum - but it wasn't. That's a working version of his mechanical calculator, the Difference Engine (No 2) - not a computer. There's also an occasional tendency to hyperbole. 'I belong to the first generation to design, build, program, use and understand computers,' says Oliveira. That would make him of Alan Turing's generation - but the author doesn't look over 100 in his photo.
While the speculative part of the book (by which I mean all the AI stuff, not just the chapter on the Singularity labelled Speculations) is very interesting, it can be quite dismissive of others' views. Oliveira seems to have no time for Good Old Fashioned AI (he should have read Common Sense, The Turing Test and the Search for Real AI) and dismisses Roger Penrose's ideas of a quantum component to consciousness as making him an 'undercover dualist', which they surely don't.
Overall, then, fairly plodding (certainly no romping) in the introductory sections, but worth reading, if you are interested in AI, for the later sections and their stimulating ideas.
Hardback:
Kindle:
Using these links earns us commission at no cost to you
Review by Brian Clegg
Comments
Post a Comment