There are two current branches of AI development.. and most of the ones listed above fall into the first branch - simulated AI. It\'s not real AI, it\'s done by software, and by writing many sweeping algorithms that a robot can follow, and will follow based upon heirerarchical order. It\'s basically very advanced programming, and it can allow something to interact in a situation it\'s never faced before, but at the end of the day, it will fall back upon it\'s core programming.
The second branch of AI, true AI, involves simulating a human brain by making neural networks... This is the field my father is working in.. but that field is barely in it\'s infancy. True, learning networks are basically simulated brains.. but currently (and for the forseable future) running in a simplified manner, and with at most a few dozen connected neurons (compared to our thousands of connections in a cluster, with millions of overall neurons/clusters). True AI has a bunch of problems to overcome - input must be carefully tailored to the type of network created, output cannot necessarily be explained (and certainly can\'t be explained by the neural network), depending on the task, if a cluster learns something \'wrong\' then it\'s output can be forever thrown off.
basics of true AI