For decades Go was the game artificial intelligence couldn't crack, too intuitive and combinatorially vast for brute-force computing. This film follows the DeepMind team that built AlphaGo, a program trained on millions of human games and years of self-play, as they prepare to challenge Lee Sedol, one of the greatest Go players of his generation, in a five-game match in Seoul. Cameras stay close on the engineers watching their creation's moves in real time, unsure themselves what it will do next, and on Sedol as the match turns into something more personal than a technical demonstration. Move 37 in game two becomes the film's turning point, a play so unconventional that commentators initially call it a mistake before realizing it reshapes the entire board. The film tracks the mounting tension in the human team, the shifting mood in the press room, and Sedol's own account of what it felt like to lose to a machine that had learned to play a game no one taught it to win this way.