AI won’t steal your job, but a centaur might

Thursday 13 August 2015

Tags:

We’re now all too familiar with the notion that the robots are coming to steal our jobs, but lets look around for a moment and see if we have any useful precedents…

We’re now all too familiar with the notion that the robots are coming to steal our jobs, but lets look around for a moment and see if we have any useful precedents… like for example when AI rocked up on the chess scene.

Way back in 1997, IBM’s Deep Blue beat the world champion Gary Kasparov – so humans gave up chess, as AI had stolen it away from us and there was no point playing any more… except, that wasn’t the case.

Kasparov realised that Deep Blue had instant access to a massive database of chess moves, and anyone could play better chess with that at their side. If that ability was fair for the AI, then why not for the human? With this in mind, Kasparov create the idea of ‘freestyle’ chess matches, where the human player (or team) can also make use of a move-database.

Within these freestyle matches, Kasparov advocates playing as a ‘centaur’ (a chess centaur being a half human, half computer collaboration) and you might be surprised to learn that as it stands today, the AI + human centaurs beat pure AI players more than half of the time. The best freestyle chess opponent today is not Deep Blue, but a centaur team named ‘Intagrand’.

The potted ‘freestyle chess history’ above should give us all two things to ponder. Firstly, that a computer alone cannot beat a human at chess (when they are allowed the same tools). Secondly, and more relevant to the coming ‘AI driven workplace revolution’, if centaurs work for chess, and are better than a human-beating AI – why wouldn’t they work for your job?