AI is just making connections that are not easy to see because of the huge amount of data that needs to be processed. It seems like magic - because we can not easily explain the path to the conclusion that the machine got - but we can describe the layers.
My college days were actually in computer science - not software engineering - which is really a mathematical field (with software engineering being applied computer science) - so, I tend to simplify thinking about software as thinking about algorithms - an algorithm is a set of steps that take us from one point to another, that's all it really is. We all know the simple ones, and software designers and computer science researches can dissect the classic algorithms for large amounts of data, but the size of the algorithm itself (the number of steps) is usually limited - as our minds just can not hold algorithms that take tens of thousands steps to get a conclusion. A machine - can, and at the end of the day - that's really what AI is (simplified form, of course).