It's official: Google's robot AI is better than human beings at Go

Google's DeepMind AlphaBot takes series 4-1

Google's AI-powered Go player AlphaBot has beaten flesh-and-blood professional Lee Se-dol for the fourth time, wrapping up the series of matches 4-1 and proving that we don't stand a chance at board games when the robot revolution comes.

Lee Se-dol managed a consolation win over the weekend but really it was one-way traffic: the AI bot rushed into a 3-0 lead after the competition began, surprising even its own developers at its Go-playing prowess.

For the uninitiated, Go is an ancient Chinese board game considered something of a high watermark for artificial intelligence to tackle: it needs an advanced level of intuition to play, not just a big bank of possible moves (there are more potential positions than there are atoms in the universe).

Do not pass Go

The fact that AlphaBot has won so comprehensively is a sign of another significant step forward for the AI tech. It's been programmed by Google's DeepMind company, dedicated to researching how neural networks can allow computers to mimic the way human beings think.

The victory means AlphaBot takes home the US$1 million prize fund, but as an AI-powered computer program doesn't have anything to spend money on (not yet anyway), the cash will be donated to charity.

If you want to relive the action move-by-move, all five games are available on YouTube. "One of the most incredible games ever," tweeted DeepMind founder Demis Hassabis after the fifth match. "To come back from the initial big mistake against Lee Se-dol was mind-blowing!"