The most obvious domain for creating an Artificial Intelligence is chess. The ability to play a game without human intervention is a goal which can be solved by today's computer easily. Most chess engines have a higher ELO score than their human counterparts.
A subpart of computer chess are pattern databases for recording game logs of played games in the past. One of them is SCID  which can be utilized for machine learning analysis. It allows to predict the outcome of a game without calculating all the steps until it's real end. This helps to increase the performance of a chess engine.