It *could* in theory be trained to filter for ideological preferences but it'd have terrible precision and cause a host of problems completely unrelated. So while you might be right on what got mistakenly marked as spam in this instance, it's nothing more than a coincidence.