How are computers programmed to play chess?

Chess: Artificial intelligence ends human dominance

You have to get a little bit afraid of the new balance of power. Whether the - for humans - almost unbelievable speed with which an artificial intelligence called AlphaZero has independently developed into an unbeatable genius in chess within just four hours.

While previous chess computers and highly developed programs have been fed and upgraded for years with the games and moves of old masters and grandmasters, in order to make use of the fund of human knowledge - or, if necessary, to put it together new and better, the new AlphaZero program has got what it takes created out of its own artificial intelligence.

In a sense, it is a chess machine with a mind that is superior to human intelligence.

The developers of the Google company DeepMind only showed their new software the basic rules of the game of chess. The AI ‚Äč‚Äčthen taught itself how to win unbeatable - without any previous knowledge - within just four hours.

AI program from the electronic super brain

Thanks to self-learning algorithms and immeasurable computing power, this new artificial intelligence has repeatedly played against itself - and in the process has developed into an unbeatable colossus, an electronic mastermind that is superior to all humans and common computers.

Chess luminaries such as the Russian ex-world champion Garri Kasparow, 54, or Peter Heine Nielsen, the advisor to the Norwegian global champion Magnus Carlsen, 27, who has been in office since 2013, are frozen in humility, respect and bewilderment about the achievements of AlphaZero.

The Google AI has just defeated the world champion chess program Stockfish, which is considered to be almost unbeatable - of the 100 scheduled games, AlphaZero won 28 games, 72 ended in a draw. AlphaZero has not lost a single one.