AI for the board game Diplomacy

As part of a larger research effort, DeepMind scientists have developed the AlphaGo machine learning algorithm, which was trained to play Go against itself without rules or pre-existing knowledge. The resulting program – known as AlphaZero – has been praised for its advanced artificial intelligence capabilities and is believed to be capable of mastering more complex games such as chess and Go in a matter of days. While the program’s abilities have excited both academics and gaming enthusiasts, the team behind it has faced criticism for using game theory to train the algorithm and the lack of transparent information about its training process. Nevertheless, researchers believe that AlphaZero will pave the way for new AI applications in games like Go, chess, and Atari.