Is Writing Style Sufficient to Deanonymize Material Posted Online?

33 Bits of Entropy ^ | 20 Feb 2012 | Arvind Narayanan

[FourPeas]

We’ve thrown open the doors for the study of writing-style based deanonymization that can be carried out on an Internet-wide scale, and our research demonstrates that the threat is already real. We believe that our techniques are valuable by themselves as well.

The good news for authors who would like to protect themselves against deanonymization, it appears that manually changing one’s style is enough to throw off these attacks. Developing fully automated methods to hide traces of one’s writing style remains a challenge. For now, few people are aware of the existence of these attacks and defenses; all the sensitive text that has already been anonymously written is also at risk of deanonymization.

Posts on Free Republic. Could they "deanonymize"?

[Larry Lucido]

Um....maybe?

[Joe 6-pack]

Protect your privacy: Plagiarize!

[TexasCajun]

Who wrote Dreams Of My Father?

[Joe the Pimpernel]

Let’s deanonymize “Dreams from my Father” and see which Chicago silver ponytail crawls out from under the rock.

[Joe the Pimpernel]

This calls for an anonymizing feature much like spell-check, that randomly substitutes synonyms for all the words in everything you write.

Sort of like babelfish only the output language is the same as the input language.

Or you could just bablefish English to German and back to English again!

Anything resembling a coherent sentence strictly coincidental.

[Paladin2]

"Could they "deanonymize"? "

Absolutely.

[Flycatcher]

Consider two words that are nearly interchangeable, say ‘since’ and ‘because’. Different people use the two words in a differing proportion. By comparing the relative frequency of the two words, you get a little bit of information about a person...

Some use the words correctly. Others don't.

"Since" refers to time:

"Since Obama was elected, America has gone to Hell in a handbasket."

But "because" doesn't refer to time:

"Because Obama was elected, America has gone to Hell in a handbasket."

Different meanings.

Uh, oh. I've just realized I've been deanomyz... deanonymousi... de-anyonim................... outed!

Yikes!

[FourPeas]

Friends and I used to do that. Humor ensued.

[Jack of all Trades]

...First they came for the verbose, pedantic, pricks but I didn't speak out because I wasn't...

oh crap

[Paladin2]

This requires an anonymous function like spell-check, write the randomly substitute synonyms for words in all that they.

A sort of babelfish only the language of the output is the same as the input language.

Or you can simply Bablefish German to English and back again to English! Such thing as a coherent set of randomly clean. Babylon 9

[Engraved-on-His-hands]

One probably can identify posters based on linguistic analysis. But I don’t think that it is as easy as one might think at first glance.

I can remember a few years back trying to identify some liberal posters based upon rather peculiar phrases that they used. I was unable to do so. It could be that some of the more “sore thumb” type of posters aren’t actually famous enough to warrant a lot of results in a google search.

I don’t use my real name. I’m not exactly famous either. But there are a lot of people who post online who know who I am. Others, with a little bit of detective work, could easily detemine my identity without going into linguistic analysis.

I used to be concerned a little about that. No longer. It’s kind of liberating. Plus, it’s helpful to understand that not that many people really care except those who love you (plus the few who hate you).

[DuncanWaring]

It deanonymized Ted Kaczinski (sp?).

[blueunicorn6]

This doesn’t work all the time. I was reading a story and the phrase “I’d hit it” came up. I naturally assumed it was LAZ, but, it turns out, it was Derek Jeter talking about baseball. Unless LAZ is Derek Jeter?????????

[dirtboy]

Unless LAZ is Derek Jeter?????????

LAZ is actually Mike Tyson. Watch your ears.

[Hunton Peck]

Seems like, if deanonymization could be refined to be, in effect, a linguistic fingerprint, it would also be possible to smear someone by “translating” self-incriminating text into his distinctive style.

[FourPeas]

Interesting point. A more creative way to forge documents.

[Travis McGee]

I assume that every key stroke I have made since Day One is fully searchable and attributable, and that I might someday be grilled under bright lights about this or that post I once made.

To do otherwise, is foolish IMHO.

[CodeToad]

“I assume that every key stroke I have made since Day One is fully searchable and attributable”

Yep. Nothing on the Internet, banking systems including checking accounts and credit cards and loans of all kinds, or anything medical or work related is anonymous anymore.

[Colinsky]

Oh yeah? Deanonymize THIS!