There are many different levels at which one can criticise AI. These vary from details of a particular method or model, up to (for example) grand criticisms of the kind that Searle, Dreyfus and Penrose make.
I'll start with a list of criticisms at different levels, to demonstrate the variety involved. Then I'll discuss Searle, Dreyfus and Penrose in more detail. I shall put a big bibliography of philosophical articles, by Chalmers, into the AI box at some stage, with lots of references to such matters.