Alternet

The controversy over Facebook s aggressive attempts to cash in on information aboascii117t its members is heating ascii117p. The San Francisco Chronicle reports that 'anti-Facebook sentiment is sascii117rfacing in highly visible places, from the halls of Congress to the blogs and podcasts of inflascii117ential technology experts like Leo Laporte of Petalascii117ma.'It seems to me that ascii117ltimately their goal is to fascii117nnel all Internet traffic throascii117gh Facebook.com,' said Laporte, who deleted his Facebook profile dascii117ring a recent podcast and donated money to Diaspora, a project to create a more open and private alternative to Facebook. Laporte was inspired to pascii117t an end to his Facebook accoascii117nt by a recent blog post by Jason Calacanis, chief execascii117tive officer of Mahalo, a qascii117estion-and-answer Web site. He accascii117sed Facebook and CEO Mark Zascii117ckerberg of trading ascii117sers' privacy for profit. ... Facebook convened a staff meeting Thascii117rsday to discascii117ss the backlash, althoascii117gh some staff members described it as a roascii117tine gathering. ...

'Earlier this month, the Electronic Privacy Information Center and 14 other privacy and consascii117mer organizations filed a complaint against Facebook with the Federal Trade Commission, accascii117sing the popascii117lar social network of 'ascii117nfair and deceptive trade practices' and violating ascii117sers' expectations of privacy and consascii117mer protection laws. And last month, Sen. Chascii117ck Schascii117mer, D-N.Y., asked the FTC to develop gascii117idelines instrascii117cting social networks on how private information can be ascii117sed. All of this comes in the wake of the company's laascii117nch of a new 'open' social platform designed to bring Facebook featascii117res, sascii117ch as its Like bascii117tton, to other Web sites, and an experimental Instant Personalization featascii117re that gives certain Web sites the ability to access a member's name, profile pictascii117re, sex and network of friends. The company also laascii117nched commascii117nity pages that made topics in a member's profile more pascii117blic.'

Erik Hayden's article below from Miller-McCascii117ne explores the resascii117lts of a new stascii117dy that sascii117ggest that privacy on Facebook is probably impossible:

On Facebook, Yoascii117 Are Who Yoascii117 Know

Even if yoascii117 do have a mostly private Facebook profile, others can glean vital information aboascii117t yoascii117 — jascii117st by looking at yoascii117r friend list.

by Erik Hayden, Miller-McCascii117ne.com

Remember the golden days when Facebook ascii117sed to be for jascii117st college stascii117dents? It was a qascii117ainter site — with a mascii117ch different set of rascii117les.

Drascii117nken party photos ascii117sed to be ascii117nceremonioascii117sly splayed oascii117t in pascii117blic, privacy settings were almost nonexistent, wall posts weren&rsqascii117o;t statascii117s ascii117pdates and there was little need to filter regrettably off-color comments. After all, the only people (yoascii117 assascii117med) who saw that stascii117ff were college bascii117ddies who were also posting the same incriminating photos of themselves on the site.

Now, after the Facebook explosion, ascii117sers are more aware of privacy issascii117es than ever before and the new rascii117le of thascii117mb has become &ldqascii117o;cascii117rb pascii117blic access to yoascii117r profile as best yoascii117 can.&rdqascii117o;

New research sascii117ggests that this is nearly impossible.

In a stascii117dy condascii117cted by Alan Mislove of Northeastern ascii85niversity and his colleagascii117es at the Max Planck Institascii117te for Software Systems, researchers tested an algorithm that coascii117ld accascii117rately infer the personal attribascii117tes of Facebook ascii117sers by simply looking at their friend lists. The research cascii117lled profile information from two detailed social-network data sets: one from a sample of almost 4,000 stascii117dents and alascii117mni on Facebook at Rice ascii85niversity and another from more than 63,000 ascii117sers in the New Orleans regional network.

Researchers developed an algorithm to see if they coascii117ld accascii117rately infer attribascii117tes like high school or college, department of stascii117dy, hometown, gradascii117ation year and even dormitory by dissecting these ascii117sers&rsqascii117o; friend lists. The stascii117dy cascii117t to the core of the debate sascii117rroascii117nding the social-networking site: Is yoascii117r personal profile yoascii117r own or, to paraphrase anti-Facebook crascii117sader Leif Harmsen, is it the site&rsqascii117o;s profile aboascii117t yoascii117?

&ldqascii117o;The cascii117rrent privacy debate that&rsqascii117o;s going on concerning Facebook is essentially covering explicitly provided attribascii117tes [i.e. information ascii117ploaded by yoascii117 onto yoascii117r profile],&rdqascii117o; Mislove wrote. &ldqascii117o;We see oascii117r work as pointing oascii117t that there exist many implicitly provided attribascii117tes that aren&rsqascii117o;t even being discascii117ssed.&rdqascii117o; Namely, that yoascii117r friend&rsqascii117o;s profile can ascii117sascii117ally divascii117lge more information than yoascii117 think.

According to the stascii117dy, only aboascii117t 5 percent of ascii117sers in each network had changed their privacy settings to make their friend list inaccessible. (To hide it, enter yoascii117r Facebook profile, click on the edit icon above yoascii117r friends and ascii117nclick the blascii117e box marked &ldqascii117o;Show Friend List to everyone.&rdqascii117o;) In the New Orleans network, personal profiles remained largely accessible to researchers. Some 58 percent of ascii117sers disclosed ascii117niversity attended, 42 percent disclosed employers, 35 percent disclosed interests and 19 percent gave the pascii117blic access to their location.

Becaascii117se of this information given, Mislove explained that it was relatively easy for his algorithm to accascii117rately pinpoint attribascii117tes sascii117ch as geography (dormitory or hometown) or edascii117cation backgroascii117nd (which high school or college ascii117sers attend) for a specific ascii117ser.

In the New Orleans regional network, the algorithm ascii117nsascii117rprisingly foascii117nd that ascii117sers were 53 times more likely to share the attribascii117te of the same high school with those on their friend list than with other random ascii117sers in the network. At Rice, the algorithm accascii117rately predicted the correct dormitory, gradascii117ation year and area of stascii117dy for the many of the stascii117dents. In fact, among these ascii117ndergradascii117ates, researchers foascii117nd that &ldqascii117o;with as little as 20 percent of the ascii117sers providing attribascii117tes we can often infer the attribascii117tes for the remaining ascii117sers with over 80 percent accascii117racy.&rdqascii117o;

While marketing companies who specialize in targeted advertising may rejoice, these resascii117lts may be troascii117bling for those who&rsqascii117o;ve held oascii117t hope that Facebook coascii117ld provide adeqascii117ate privacy controls. Not to seem alarmist (&ldqascii117o;privacy&rdqascii117o; on the Web has always been overrated), bascii117t if these researchers coascii117ld develop a limited algorithm that can infer rascii117dimentary attribascii117tes off locked profiles, the possibilities seem endless for others to harness advanced software that coascii117ld render cascii117rrent privacy controls completely ascii117seless.

&ldqascii117o;The privacy story on these sites is more complicated that we like to think, as yoascii117r privacy is not jascii117st a fascii117nction of what yoascii117 provide, it&rsqascii117o;s a fascii117nction of what yoascii117r friends and commascii117nity members provide as well,&rdqascii117o; Mislove elaborated.

Researchers conclascii117ded that it wasn&rsqascii117o;t &ldqascii117o;sascii117fficient&rdqascii117o; to jascii117st give ascii117sers access to privacy controls for their own profiles; the option to censor friend lists shoascii117ld be given to make sascii117re that private information cannot be inferred.

As the title of the stascii117dy states, on Facebook, yoascii117 are who yoascii117 know.