An AI (at least assuming it results from foreseeable extrapolations of current research) isn't going to have intrinsic motivations like "boredom" or "happiness"

An AI (at least assuming it results from foreseeable extrapolations of current research) isn't going to have intrinsic motivations like "boredom" or "happiness" so anthropomorphic thinking a very useful way to reason about its behavior. Its only "motivation" will be to maximize the objective function it has been given by those who (at least nominally) control it. If the AI is very intelligent that still allows for a great deal of unanticipated behavior though.