I think that artificial intelligence is related to, but distinct from, machine learning. Machine learning, at least as it is practiced today, comprises automatic methods to learn from large volumes of data against rather structured problems: given data in a very exactly-laid out format (a feature vector, a graph, a tokenized corpus, an image), make an inference about one specific quantity (e.g. probability of class membership).
People can bicker about what, exactly, AI is, but I think it's fair to say that AI aspires to be more general than simply ham/spam e-mail classifiers or recognizing handwritten digits. I think the dream among AI researchers is that AI will be generic as self-teaching systems which can apply knowledge gained in one area to a new area, or to be reasoning machines, or work with non-structured data on open-ended tasks, or be creative, or other lofty goals.
That said, I'd be surprised if the tag artificial-intelligence is used in that sense on stats.SE; I suspect (and even hope!) that questions about AI in the sense that I describe above are self-sorted to the AI.SE website, as that would appear to be their unique value-add to the SE ecosystem.