First, a GET request is in no way a statement that the user intended to view an ad: is accidentally clicking an ad fraud? Of course not.
Second, a GET request is not even an indication that the user clicked on some element. It's not even a statement that the user requested a particular resource. Indeed, when the ad itself is requested you could hardly call that user-requested; if your analysis were correct I see no reason we couldn't extend it to say that ads themselves constitute a fraudulent statement that "this content is something you requested".
Third, even if a GET request amounts to a statement of user action (if not intent), adnauseam's actions do not constitute a false statement of user action/intent. There are lots of ways to activate the default onclick handler; physically pressing a mouse button is only one of them. Trackpads and touchscreens are other ways; are they fraudulent? Keyboard shortcuts and other keyboard navigation are yet more, are they fraudulent? No, of course not, since they accurately represent user actions. Installing adnauseam is just another method of performing the same action, clicking on ads. It accurately implements the intent of the user, more so than clicking, which can be accidental.
IANAL, but based on your own criteria, without a false statement of material fact there is no fraud.