> It is understandable, expected, and desirable that artificial intelligence systems provide accurate and factual information about empirically verifiable issue

> It is understandable, expected, and desirable that artificial intelligence systems provide accurate and factual information about empirically verifiable issues, but they should probably strive for political neutrality on most normative questions for which there is a wide range of lawful human opinions. Well, see, that's just the problem, because many political stances (and arguably, somewhat more conservative stances) are clearly not empirically defensible. Take the stance against gay marriage, for example. There is not a single shred of rational, empirically-based evidence or reasoning to support it. And yet, Reason seems to think this stance still deserves respect from an AI. I disagree.