Two weeks ago, AOL mistakenly released about 3 months worth of search logs. In the data, the user-ids were replaced with numbers, in an attempt to “anonymize” the data. Despite this anonymization, the release of this data clearly violated the privacy of a great number of AOL users, and AOL quickly stopped publishing the data, but by then it was too late. The cat was out of the bag.

The data can still be downloaded via BitTorrent, and you can sift thru the data using websites like this one and this one.

Other websites started publishing amusing and bizarre excerpts from the logs.

The data in these logs is fascinating and has already taught me a few things.

1) Most AOL users are noobs. Okay, I already knew this. I mean, c’mon: all the grandmas and grandpas in my family stopped using AOL over two years ago, so clearly anyone who is still using it has got to be pretty clueless (sorry grandma!). The logs make this pretty apparent: they are chock full of stuff like this:

2) AOL search isn’t very good. Okay, I knew this already too. The number one thing that AOL users type into their search box is the name of a better search engine. Google is #1, and Yahoo is #2. Clearly, these users are looking for something better, but just don’t quite know how to obtain it. Also there are are a very high number of searches in which the first 7-8 results are deemed to be insufficient, and the user continues looking.

3) “Anonymizing” web logs by using numbers of instead of user-ids does little to protect the privacy of the users. The New York Times very quickly sussed out the identity of one of these users, and I am sure others will follow. A search of phrases like “my password” reveals all kinds of potentially damaging stuff.

4) Anonymized search logs, in addition to not protecting privacy very well,are highly entertaining. It is incredibly easy to find search histories that read like dime store pulp fiction. Consider the following excerpts from user 14162375, which I dug up by searching for the phrase “kill my”.

A good way to find salacious material in this database is to search for phrases like “clear all” or “erase history” which almost always follow searches the user would prefer to cover up.

Anonymous User 4320454, is one of a handful of users who attempted to clear their search history by typing “clear all searches”, in this case because he had been searching for “how many days or hours for cocaine to leave the system to be clean for drug urine test for employment”. It didn’t take.

Ohio user 4838451, interrupted a long night of searching for Fran Drescher porn to search for Senator Mike Dewine, followed by a few more porn searches and then “clear all searches”. It didn’t take.

This entry was posted on Tuesday, August 15th, 2006 at 5:03 pm and is filed under Good reads, Linkydinks. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.