AlphaGo made the famous “Move 37” during its 2016 Go match against Lee Sedol.
In the second game of the five-game match, played in Seoul on 10 March 2016, AlphaGo placed a stone that commentators initially considered highly unusual. The move was later recognized as strategically creative and changed the direction of the game. AlphaGo won that game and ultimately defeated Lee Sedol by four games to one.
AlphaGo was developed by Google DeepMind. Its system combined neural networks with tree-search methods: one network estimated promising moves, while another evaluated positions. It had been trained with human games and then improved through self-play.
“Move 37” is often described as move 37 by AlphaGo, although the exact move numbering can vary with notation conventions. The match mattered because Go has a vastly larger search space than chess, making brute-force approaches much less effective.