I would imagine the list of things that will count against you in those situations will always be much longer than the list of things that can count for you, i.e. just because A can count against you doesn't mean that ¬A will count for you; ¬A will just be 0 points and the assumed baseline.
In one sense, it's a zero-sum game. If person A is a better debtor, A gets a better rate than B. However, a more sophisticated portfolio will try to find anti-correlated debtors. Rather than lending to 100 low-risk people all similar to each other, I'd rather diversify and include some very different people that may be higher risk on an individual basis; the aggregate risk can be much lower than individual. To that end, the more the lender can learn about me, the more chance I have to reveal how anti-correlated I am with others.
There is plenty of questionable shit happening right now in the NSA and stuff. I don't see why we need to resort to fearmongering about absurd hypotheticals.