Some days ago AOL, or at least a team within, decided to release the search dataof more then 650,000 users. They did replace actual user names with random numbers. Using those numbers you could still track all the search terms of a single user.
Then this announcement came: “A Face Is Exposed for AOL Searcher No. 441774“. By using the search terms they were able to narrow down to a single person.
This makes us wonder how much information we are leaving behind, even anonymously, that allows others to uniquely identify us.
Recent attempts at creating an Internet Meta Identity System (see Infocards and others) do include the possibility to identify yourself more anonymously (for instance being able to prove you are over 21 years) without revealing your identity. However, most sites will still enable tracking cookies. So, over time, they might be able to identify you.
For some reasons I am afraid it is impossible to design a system in which none of the participating parties, except the user, can accumulate enough information to uniquely identify someone. Knowing this, should be part of the user education. Upcoming meta identity systems will enable a smoother and more powerful experience for both the end user and relying parties but it will not completely protect your identity or privacy. You just leave too much footprints behind to ensure that.
While writing this, I saw this Google announcement telling they will still store search data, despite the potential privacy concerns.