It isn’t strictly accurate to use “artificial intelligence” to refer to currently available technologies like robots, autonomous vehicles and digital assistants that implement machine learning. The algorithms that drive these technologies are still fairly constrained to the problems they are given to solve. In other words, they aren’t yet able to independently collect data and derive new models without external programming.
The AlphaGo program, developed by Google, is an exciting step forward in artificial intelligence. It was able to beat human players at a challenging game, a feat that had previously proven impossible. But, AlphaGo is still engineered to do one thing very well, it is not an autonomous intelligence that could, for example, say, “Hey, can we switch it up and play checkers instead?”
While we are making some important and exciting steps in the direction of artificial intelligence, we still have not reached the science-fiction version of AI.