Google Researchers Warn of Automated Social Info Sharing

holy_calamity writes "Researchers from Google have written a paper about how social networks can undermine privacy. The most interesting scenario they discuss is 'merging social graphs' — when correlating multiple social networks makes it possible to reveal connections that a person has intentionally kept secret (PDF). For example, it may be possible to work out that a certain LinkedIn user is the same person as a MySpace user, despite their attempting to keep their profiles separate. The Google solution is to develop software that screens new data added to a social network, attempting to find out if it could be fodder to such data mining."

Comments

Re:Other people may publish information about you

[Creepy Crawler]

--While that is a completely fair thing to point out, there is a very important thing that it misses: other people can put information about you online, without your permission, and that information is just as subject to analysis as what you put up.

Absolutely true. However, this information also falls under libel laws in cases which it false and harms the subject.

---The two best examples that come to mind right away:
---1. Facebook allows users to tag photos with the names of the people who appear in them.

It may be a stretch to call it this, but posting stories and pictures usually is a form or journalism, regardless the content and methodology of dissemination. Also, photographers need a release whenever the photo is not taken in a public area, or accessible from a public area of a person who does not consent. It is not a crime to fail to get a release, but a nice tort claim.

---2. Google Street View puts photos of your residence without asking you for permission, and correlates it to a bunch of other stuff like geographic information, satellite images, yellow page listings, web search results, etc..

Google, as long as they obey the law in terms of public/private property, they have full legal standing, and shouldn't be regulated. However, they did get in trouble when they went down private roads, with nice posted NO TRESPASSING signs peppered all over, which the GooCam dutifully captured.

---Notice that both of these acts are perfectly legal, and while the second arguably should be regulated and restricted by law (the aggregation, correlation and publication parts, not the picture-taking part), the first one ought not to.

Sounds like Google-Hatred. Somehow we should just make a law for Google, because they spy on us!! Guess what: You have as much access to the same pictures as I do. And if there is any sort of law, it's that I'd like posting on who (specific names/residences) viewed some area in high resolution. You know, watch the watchers. Sous-veilance.

For me...

[Anonymous Coward]

I don't have a LinkedIn profile. I had a MySpace, but I killed it. My Facebook account does have political rants and opinions, but I don't have any pictures that could prove embarrassing. Google my name, and you'll find political rants I've posted on Yahoo! Groups and similar.

However, I currently have something that can't hold a candle to being careful on Facebook: an employer that doesn't care what I do off the clock. As long as I stay out of trouble with the law and notify my employer if I'm moonlighting, all they care about is that I show up for work and do my job. Should I get on-call duties in the future, I'll be held to a written SLA, and as long as I can respond to calls within an allotted time, they don't care what I was doing. I can even apply for other jobs without fear.

I used to be a K-12 teacher--if you ever think you're being spied on off the job, just be glad you're not a teacher.

- Twice, I was indirectly threatened with termination for looking for other jobs.

- The local newspaper photographed a teacher with a group of animal-rights protesters when the circus came to town. Parents started complaining to the principal saying they didn't want their children in that teacher's class.

- Some students found a picture online of a naked woman who closely resembled--but wasn't--a high school teacher. They took the picture to the school's principal, and the teacher wound up having to defend her teacher certification against revocation--fortunately, she won. I'm not even sure if there's a rule against teachers posing naked, as long as it doesn't directly involve students.

Granted, should I ever find myself looking for a job, a potential future employer might Google me and decide that I'm too liberal (or too something-else) to work there. But the question here is: do I really want to work somewhere they hold employees' personal, legal, off-the-clock lives under a microscope?

Re:Hmm...

[vux984]

I thought Google would be the ones undermining privacy in this case...

How do you think they "discovered" this?

Yah, we know... *they* only just realised this...

[Klootzak]

They've ALREADY opened Pandora's box, now they're trying to get some publicity on how they can fix the problem they (at least helped) create.

I don't hate Google, I think they tend to go "Oh, cool, we can do this!" without thinking through the implications of it... much like other clever people in history have done.

You're missing the point

[Estanislao Mart�nez]

I'm not sure I understand why a picture of your house, taken from a public road, constitutes an invasion of privacy.

But that's the whole point: the picture of your house, taken from a public road, as one isolated token of information, does not posit much risk to your privacy at all. Tons of individual people in our society own cameras, and take photos in public places that depict other people's property, and everybody agrees that the owners of said property should not have a right in general to prevent others from taking such photos. The privacy laws we have are built to protect individuals' rights in that sort of isolated case.

The problem is when a corporation starts taking such photos systematically, aggregating them all together and correlating them with other systematic data sets. In that situation, a photo that just happens to contain your house is no longer just that; it's a piece of information that can be used to access many other pieces of information that may allow somebody to infer facts about you that you would rather prefer they couldn't.

Google Street View is only the start. Just wait for the day when digital cameras commonly include GPS units and automatically tag each photo with a precise location and time, which can then be cross-indexed with a geographical information system like Google Maps. I can imagine it already: your wife carelessly forgets to close the window shades one day when they're changing. A neighbor takes a photo of her naked, and posts it to 4chan. Thousands of folks copy the photo all over the web. The photo has GPS information in the EXIF tags. Creepy /b/tards start stalking your wife. You give a resume to a potential employer with your residential address in it; they look up the address in Google Maps, click on the link to show image search results taken nearby, and are treated to a naked picture of your wife.

That's an example where the photo in question is probably illegal to take, but other examples may be concocted where the picture, by itself, is fine. The point of the example isn't the photo; it's how the technologies that we have today for associating one item of information to others make it too easy for people to find out more about you than they should be able to.

To sum up, the privacy laws we have today are laws that were designed to protect people's privacy in yesterday's, pre-computer world. Because of this, they primarily address things like whether somebody had the right to take a given individual photo, and not whether somebody is empowering others to infer facts about you by correlating many individually innocuous items of information.

Re: don't need fancy data-mining tools for that

[bill_mcgonigle]

However, anything you do tell them can (and will) be used against you.

Yes, and only against you.

(bug #8332 has been filed).

Re:So when do we give up on privacy?

[martin-boundary]

Not necessarily. While it is true that once some piece of information is on the internet or on someone's hard drive, then it can be assumed to never go away, and eventually be collated, there is in fact a theoretical way to neutralize the big brother. It's known as spamming, or poisoning the well, or hiding in plain view, and I expect those techniques to actually help privacy in the long run, however crazy that sounds.

The reason that Google and other companies can mine so much useful information about you is because that information is substantially true. If, for example, the information was 90% useless and totally contradictory, then it would be impossible to reassemble a correct picture about you. If this was common, then people would not usually trust what's written about you or anybody on the web, and you would regain some level of privacy.

As privacy becomes more of an issue, I fully expect SEO companies and spamming outfits to reposition themselves as privacy protection agencies. There's going to be a lot of money to be made in helping people, and the same bunch of scum who are now crucified for spamming may well be hailed as pioneer heroes in, say, 20 years.

Re: Obfuscation

[TaoPhoenix]

This may be the best common ground. The internet made "leveraging the genius" the thing of the 2nd decade (give the world another year to rest & prepare).

Now all any organization has to do is locate one of some 10,000 people worldwide with the knack at seeing ultra-patterns. Most of us aren't that interesting to bother with, so a little camoflage to avoid 5 second name searches is usually enough. But what the real point of the Mrs. S. effect is, "if you annoy the collective Net, they'll borrow a pattern expert and pulverize you".

Web 2.0 and 2.1 are about Sharing.

Web 3.0 and 3.11 will be about privacy.

Re:Regulation won't solve this problem.

[TaoPhoenix]

All. Everyone will have everyone's information. There's no way it can be removed from everywhere. We learned that if you try to nuke the top ten sites, it just floats around the 100 second tier sites you'll never find.

All we can do is resign ourselves to it and be boring enough not to really be worth the time.