“Traditional” AI grew out of efforts to crack enemy codes in the Second World War. It aimed to capture human intelligence by following vast lists of rules programmed into a computer. Today, this approach is best known for creating Deep Blue, the computer that beat the chess world champion Garry Kasparov in 1997.

But this strategy has serious limitations because it seems unlikely to produce anything that really resembles human intelligence. Instead, a new wave of AI is slowly making its mark. It relies to a large extent on coaxing complex behaviours from the interaction of simple components.

So, for example, networks of artificial brain cells can learn and recognise patterns. Already such neural networks are advising financial wizards about investing their money and helping doctors to diagnose cancer.

More about the death of AI here