AI further reading suggestions.

This page contains some suggestions for further reading on specific aspects of the course. It is obviously not intended that you read all these books! If you wish or need to follow up an aspect of the course, these may help.

Note: I quote some prices below, but these may not have been checked recently! Prices are all in pounds and are given for the paperback edition if available both paper- and hardbound. All of these prices were checked on John Smith's search engine on 26th January 1998. Most of the books on this page were in stock and all were orderable.

Philosophy of AI

When was the last time you were recommended a book costing 2 pounds? An introduction to Turing's work is contained in Andrew Hodges' 'Turing' (Orion, 2.00) in the Great Philosophers series. Hodges also wrote the definitive biography `Alan Turing: The Enigma of Intelligence' (8.99) and hosts a set of web pages devoted to Turing.

There are many books covering background and source material on the philosophy of AI and whether machines can think. An example is `The Mind's I' edited by D.R. Hofstadter and D. Dennett (Penguin, 15.00) which contains reprints of the paper by Turing on the Turing test and the paper by Searle on the Chinese Room thought experiment and much else.

A number of books have been written on the whole AI enterprise, ranging from very positive to very negative viewpoints. Daniel Dennett has written `Consciousness Explained' (Penguin, 9.99). Searle gave a series of Reith Lectures in 1984 on his thought experiment and published these as `Minds, Brains and Science' (Penguin 6.99). Other examples worth looking at are `Computer Power and Human Reason' by Joseph Weizenbaum, (Penguin, 8.99), `What Computers Still Can't Do' by Hubert Dreyfus (MIT Press, 14.95), `Artificial Intelligence: The Very Idea' by John Haugeland (MIT Press, 14.95), and not one but two books by the physicist Roger Penrose: `The Emperor's New Mind' (9.99) and `Shadows of the Mind' (7.99). Roughly speaking, Dennett is the most positive towards AI, then we get gradually more negative as we go through Haugeland, Weizenbaum, Dreyfus, Searle and Penrose, although this only means positive or negative to the prospect of creating thought and consciousness through AI, rather than positive or negative towards achievements such as game-playing or speech recognition.

An all time classic in AI is the 1963 collection of essays `Computers and Thought' edited by Feigenbaum and Feldman. This has recently been republished (MIT Press, 15.50).

No AI reading list is complete without a pointer to Douglas Hofstadter's `Godel, Escher, Bach: An Eternal Golden Braid' (Penguin, 17.99 paperback) although the book ranges much further afield than just AI. Much less satisfying is Hofstadter's later collection of Scientific American Essays `Metamagical Themas: Questing for the Essence of Mind and Pattern' (Penguin, 17.00 paperback), but it does contain Hofstadter's dialogue with the Nicolai system and the circumstances in which it took place. (Not on AI at all but very stimulating is the `Person Paper on Purity in Language' in the same volume.)

Games Playing

Jonathan Schaeffer has written a fascinating account of the development of the Chinook world champion checkers program, in his book `One Jump Ahead' (Springer Verlag, 25.50 hardback). This is a warts-and-all account by the main author of the program and probably the only book ever written for the lay reader about the development of a single computer program. Because it is for the lay reader, the AI techiques used are only sketched.

Why not Play a game with the World Champion?