Say I want to show an "ad" to white, heterosexual, protestant men with libertarian leanings, no history of depression, unlikely to use alcohol in their free time, who don't look pornography, and are predisposed towards compliance with authority currently employed in the technology industry earning over $150,000. Say my "ad" is for a high-paying job opportunity. Or housing. Facebook definitely has the ability to let me do this.
I find that unsettling, in that people can use Facebook to apply prejudices against information I'd prefer to keep hidden but that Facebook has inferred via data-mining even without knowing who I am or actually learning that information about me.
The "stalker" scenario presented here is certainly frightening (although as a heterosexual man I can only really understand the fear in the abstract), but the "reveals information I'd prefer to keep hidden to people with a financial motive for knowing it" scenario is already getting uncomfortable.
If I showed up to a job interview where someone wanted a "white, heterosexual, protestant men with libertarian leanings, no history of depression, unlikely to use alcohol in their free time, who don't look pornography, and are predisposed towards compliance with authority", it would be a strong disappointment for both of us. The problem is not targeted ads. The problem is that people suck.
But say that I am only slightly racist. If I knew you were Black, I probably would not go out of my way to say high to you. But once we are having a conversation it's easy for me to relate to you and my prejudices are more likely to be overridden.
It's easier to fix people sucking if said suckage is easily observable though...
What I'm mostly saying is that a lot of the you-are-the-product hate is often presented hand-in-hand with “facebook just gives your information away to advertisers willy nilly so they can make money”, and I'm saying that's not really accurate. Facebook need you to stay on their site, so they get you to share information because that brings you and others onto the site. For advertising purposes, though, they don't have to give away any information, really. And at that point it becomes a matter of “do I trust Facebook with my information”, rather than trusting anyone advertising on facebook with it. This is a smaller question, so it is fundamentally easier to answer.
Now, the fb API of course expands the question to “do I trust facebook… Or any site I visit that uses the facebook API”. But still, there is a “me” component there—I have to visit those sites. This is still a potential problem, no question, but it is not the oft-presented specter of facebook sharing your private information with an arbitrary third party for the sole purpose of presenting advertisements on the site. That's the refrain I hear a lot that I wanted to say something about.
And who don't look at pornography?
Facebook's whole lever for motivating people to provide that information is so they can share it with their friends. Why would anyone fill out a demographic profile to be kept under lock and key by Facebook for advertisers?
If Facebook did exactly as you suggest and walled off advertisers from all but aggregate user detail, it wouldn't have affected this app, or anyone using similar techniques.
Let's say Facebook (a) didn't offer advertisers direct personal information, only access to demographic profiles, and (b) set some minimal standard so that you can't target profiles that result in fewer than 15 people. (What's that data set that does that already?)
Can't a sufficiently large advertising aggregator run a sufficiently large set of multivariate ads, do Statistical Magic, and discover the real demographics of each of the users - at least enough to reassociate them with their real-world, extra-Facebook identity? Sure, they'd be violating ToS and probably various laws, but we're talking about attack surface, not public policy (yet).
This paper about "k-anonymity" and "privacy-preserving distributed protocol" might yield some clues, but my brain is fundamentally incapable of reading those once-"beautiful" LaTeX fonts. (I would pay good money for an app/Mechanical Turk/whatever that reprocesses text with a modern font.)
How much?
I will re-write the paper for you using the font of your choice for one millon dollars.
As an example, Google allow advertisers to target what people search for, without telling you who they are.
The users of this app are most likely to be introverted young men rather than serial killers and rapists. It's impossible to hurt someone with an app so linking the existence of any app with the dangers of stalking, rape and murder seems to be fear mongering.
You didn't understand cstross's sentence. He said that the social networks are motivating their users to disclose information [that is] relevant to advertisers; he didn't say that they are disclosing the information directly to advertisers.
2) Handwaving about applications and like buttons aside, it's pragmatically trivial through the facebook ecosystem to capture a great deal of data about individuals that they might not natively believe they are exposing. I know this because I personally do this (for benign reasons).
This is what makes most complaints in the facebook-makes-money-off-of-you vein a complete misunderstanding of technology.
For the purposes of advertising, facebook does not need you to disclose relevant information to anyone but facebook. Advertisers need never come anywhere near that information. The data doesn't even need to be sent to them in aggregate. All advertisers need to do is say “I want to target this demographic”. They then need to get no data other than “someone from this demographic clicked on your ad”. That's it! Facebook has the data, facebook does the matching, facebook provides feedback in clicks—this, yes, in aggregate—and no other information needs to be exchanged.
There is a broader argument being made here, about exposing personal information and what have you, but facebook doesn't need you to make your information public so that advertisers can get to it, they want you to make your information public so that that guy you met at a party last night can find you, friend you, and start liking your posts, thus making you feel that your posts are getting more attention, thus making you feel good about yourself, thus making you want to post more, thus making others want to get on more to read your posts, etc, etc. Of course, they also want it public so that guy over there can look you up and see some of your pictures, even if he doesn't necessarily intend on friending you. That way, said guy is also on facebook, using it more than before.
So I would argue that how public information disclosure is on facebook is far more about retention and usage than it is about advertising. Advertising can be and, at least as far as my understanding of it goes, is done without having to disclose anything to the advertiser. And that's as it should be, since the technology lets them not leak that info. Why would they want to provide that information if they can achieve the same results without doing so?
With that in mind, I wish the facebook privacy argument would stop presenting advertiser information swapping as a real issue. There are reasons that privacy is an issue on facebook, foursquare, etc., (indeed, this app is an excellent example) but I don't think advertising is one of those reasons. That can be done with very little information revealed, even in aggregate.
(It's possible I missed a detailing somewhere of facebook explicitly exposing information to advertisers in a way different from what I mention above; if so, I'd love a link and I'll sit down and be quiet now :) It's also true that between fb advertisements and installing fb like buttons, it's possible to gather more data on a user than concentrating on advertisements alone. But that's a complaint about fb giving your information to apps, and while it's a lot harder to tease out the motivations there, I think they still derive a greater benefit from that in terms of retention and usage than they do in terms of advertising.)