Machine learning is one of the most active research areas in Artificial Intelligence today. Whereas computers have proven capable of concentrating and expressing human expertise on some areas -the game of Chess for instance- this is a large engineering endeavor.
The possibilities of going beyond to new paradigms where machines can be let alone to gain expertise by themselves exploring worldly problems in more or less directed ways are open. There have been some successes and many failures.
Sometimes the effort necessary to launch a successful learning experiment is so big that in the end there is no real learning happening - The experimenter has coded all the knowledge in the form of delicate parameter tuning and custom analysis functions.