A New Way of Thinking

by Thomas Masuch

In Europe, playing chess is a tradition that goes back around 800 years. For some, it's the undisputed king – or perhaps more appropriately, the queen – of all games. Others see it as nothing more than a waste of time. Regardless of one's own personal viewpoint, it’s a fact that chess has produced no shortage of stories and myths.

Since 1996, however – when the computer Deep Blue defeated then-world champion Garry Kasparov in a game – a changing of the guard has taken place: The machines have left us humans further and further behind. These days, grandmasters no longer even consider taking on the most powerful chess programs. That’s how unbalanced the playing field has become.

As many experts have tried to learn from the superiority of these computers, chess has grown more technical and analytical. That said, a new era may have begun just a few weeks ago with Google's AlphaZero – a program that has brought something akin to human creativity back into the highest level of the game. AlphaZero is a self-learning program that was developed by DeepMind Technologies. Upon being acquired for an estimated $500 million in 2014, the company received a new name (Google DeepMind) and a new stated objective: »solving intelligence«.

Triumph of the new artificial intelligence

It took AlphaZero all of four hours to learn the game of chess by playing against itself. This past December, AlphaZero then went up against the previous year’s world- champion chess computer, Stockfish, and came away with a clear victory. In and of itself, the triumph of this new artificial intelligence wasn’t all that surprising given the tremendous processing power at AlphaZero’s disposal; it was the more the manner in which it won. The program astonished observers with creative moves that previous computers hadn't even included in their calculations. In some respects, AlphaZero's style of play even challenged our very understanding of chess up to that point. Its developers’ answer to the question of what makes it so good? The fact that the program no longer relies on human input.

This statement may be frightening; indeed, it's at least strangely exciting to watch as a new way of thinking emerges. And we're not talking about just chess, of course: In the future, what will cars, airplanes, or everyday products end up looking like when computers can learn from their mistakes and start getting creative? It seems almost certain that additive innovations will play a vital role in this process. After all, no other technology makes it so easy to turn creativity into reality.