Humanity's best Go player finally beat Google's AI for the first time (GOOG)

Google DeepMind's AI was beaten by Go world champion Lee Sedol on Sunday in the fourth game of the Challenge Series tournament.

It's the first time that Sedol has managed to beat the computer.

Google DeepMind CEO Demis Hassabis tweeted that Sedol "was too good for us today and pressured #AlphaGo into a mistake that it couldn’t recover from."

Although this is a big win for Sedol that proves DeepMind's computer can be beaten, Sedol still can't win the tournament. This is game four of a five-game series, and DeepMind won the previous three games.

"I've never been congratulated so much just because I won one game," Sedol said at a press conference following the match.

"Lee Sedol is an incredible player and he was too strong for AlphaGo," Hassabis said. "We are also very happy because this is why we came here — to test AlphaGo to its limits."

Hassabis said the DeepMind team will analyse the loss when they return to the UK to figure out which problem caused the program to lose.

Commentator and Go expert Michael Redmond said that "AlphaGo played an interesting game." But he said that Sedol played a "brilliant move" on turn 78 that took AlphaGo by surprise.

Throughout the tournament DeepMind's computer has made unusual moves that a human player typically wouldn't consider. Some commentators have wondered whether they are mistakes, but usually they've just been unusual moves.

However, The Verge reports that the AlphaGo program made one unusual move in turn 79 of this game. Hassabis tweeted that the move was a genuine mistake, rather than a rare move:

This won't be the end of the tournament — there's still one game left to play on Tuesday.

You can view the fourth match here:

Join the conversation about this story »

NOW WATCH: This trick will make reading on your iPhone at night much better