Many people have been talking about artificial intelligence for quite some time now. However, far fewer people are capable of explaining what AI actually is. So, what’s it all about? … The term Artificial Intelligence was coined by the U.S. mathematics professor John McCarthy in 1956. So the concept is by no means as new as people think. The topic captured more public attention in 1997, when the chess computer Deep Blue beat the world champion in chess at that time, Garry Kasparov. That was more than 20 years ago, and chess computers are no longer really revolutionary.
A recent example shows what AI can do today. Compared to chess, the Chinese board game Go for a long time was considered too complex to be played by computers. This complexity is due in part to the fact that the number of possible combinations of playing pieces on the board exceeds the number of atoms in the entire universe. However, in 2016 the AI-supported AlphaGo program beat the world’s best Go player by 4 to 1. This feat attracted attention. However, an even greater stir was caused by an enhanced version of this program, AlphaGo Zero, which was matched against AlphaGo, so that two computers were competing against each other. What made the enhanced program notable is that it had learned the optimal strategy not from previous games played against human players, but by playing against itself. This type of machine learning is referred to as reinforcement learning. The astounding result of this match is that AlphaGo Zero beat AlphaGo 100 to 0. This example was taken from the book Realitätsschock (The Shock of Reality) by the German blogger and writer Sascha Lobo. It impressively demonstrates the potential of AI.