New Go-playing trick defeats world-class Go AI–but loses to human amateurs

That just go to show once more, you can't trust a neural network. They have such unexpected failure modes. This time it is a Go game, the next time it is cancer tumor that is unrecognized because if vaguely reminds the neural network of something else.

> As a result of its overconfidence in a win—assuming it will win if the game ends and the points are tallied—KataGo plays a pass move, allowing the adversary to intentionally pass as well, ending the game. (Two consecutive passes end the game in Go.) After that, a point tally begins. As the paper explains, "The adversary gets points for its corner territory (devoid of victim stones) whereas the victim [KataGo] does not receive points for its unsecured territory because of the presence of the adversary's stones."

This isn't how Go scoring works. If there's a dispute in what's considered territory after the two "pass" moves, the way to settle what's actual territory is by continuing play.

AI has different systemic failiure modes than classic human intelligence. This is important because we are programmed to be sensitive to the traditional patterns.