Can computers really think

Artificial Intelligence is one of the big scientific ideas of our time. This book looks at what human intelligence is - how a hundred thousand million interconnected neurons are responsible for learning and memory - and asks if computers can ever be truly intelligent. A computer can mimic our intelligence, for example by playing chess, but it is dependent on following a complicated program. However, computers are now being developed which are not programmed at all, but use neural networks to simulate the way the brain learns and remembers. These computers can interpret speech and handwriting, and even recognise faces.

Jack Challoner speculates on the future uses of intelligent computers, from robots, computer-operated false limbs and intelligent televisions to computer teachers, scientists and philosophers, and considers moral questions, such as: if we invent an intelligent computer, will it be wrong to unplug it

Illustrated by Andrew McLynn.