MIT’s Technology Review had an interesting article about AI, Kasparov and man vs machine, written by a distinguished professor of philosophy. Unfortunately, the professor misses a huge portion of understanding how AI actually works.

(Speaking of, Gary Kasparov has an interesting article about the match from his perspective here.)

Chess is a reasonable target to which Artificial Intelligence can apply itself. AI offers some interesting and useful branch pruning concepts which help to avoid searching the full space of possible chess moves.

Go is a fascinating game that computers are not even close to defeating humans on. There are several reasons for this, but the largest one is that the search space is orders of magnitude larger than chess, and the pruning algorithms are less well suited to reducing that space.

As for the author’s comment about making better ‘brainchildren’ than our biological children, well, we can’t even envision how to do that yet. Despite the best effort of Science Fiction dreamers, and AI hype artists, Artificial Intelligence really isn’t. Some people have started to proclaim that AI should mean Artificial Insects, since advanced AI behaviors far better map to what insect behaviors than human.

There are reasons that general purpose AI’s do not exist. Complexity and flexibility are the two biggest ones. Until we can make several quantum leaps in the art of understanding HOW we think, AI will never excel to anywhere CLOSE to the flexibility of a human child.