I guess the problem isn't targeted advertising per se. It's that you don't know what your personal information will be used for in the future. Maybe it will give you better, more targeted advertising. Maybe it will be used to deny you an insurance policy or be sold to a recruiting firm that will reject your job application. I guess Facebooks terms don't allow that, but those terms can change at any time. Maybe not ne…
I do understand concerns over what happens to your data moving forward (changing laws, organizational/management changes etc.). But again, I can't think of any of the information on a social networking site being all that damning. The information that I share on say Facebook or Google+ (at least for me) never really consists of ultra-private information. And I think your example is a bit extreme.
The problem is, extreme eventually happens.
Nazi Germany has been such a commonplace rhetorical device on the Internet for so long [1] that we end up easily dismissing it when it's indeed a useful comparisson. This, for many reasons, would be one such case (we can debate that through email if you want), but let's dismiss it anyway.
The important fact here is that bucket loads of (your) personal data are being processed, correlated with each other and stored into centralized server clusters — which have become very tangible assets.
If someone ill-intentioned — a government, an interest group, a company — gets a hold of these assets — through power, through craft, through acquisition — they can use it to target you for whatever their purpose.
In a warfare scenario, that could mean targeted, granular misinformation; targeted terrorism through AI-generated blackmailing, tailored on an infividual level; mass identity hijacking. It could make war a personal matter like it has never been before.
(And on the other side, there's the Nazi Germany scenario where, unbeknownst to you, you are being blacklisted for a genocide to come — after all, the first step is classification. [2])
The fact that we're living rather peaceful times doesn't mean that a potential weapon like these databases should go unregulated; and the fact that we see no storm in the horizon doesn't mean we shouldn't care about the possibility.
As long as you can imagine what sorts of bad things could be done with that personal data, the argument is: you should care because you don't know for a fact what tomorrow will look like.
[1] If Wikipedia is to be believed, it's been 20 years now since Godwin's Law was first formulated!
[2] http://www.genocidewatch.org/aboutgenocide/8stagesofgenocide...