Google's DeepMind artificial intelligence (AI) is now taking on the 3D navigation and puzzle-solving game that teaches an ant to play soccer in which it has learned how to chase down a ball, dribble, and then score a goal. In a blog post, DeepMind's David Silver explains how the AI is capable of solving the ant soccer challenge without any prior knowledge of how to play. It builds on the AI's experience of playing Go, as well as the ability to learn similar to the way human learn -- by making judgement calls.