The Life and Times of Multivac
With apologies to Isaac Asimov, it's this server advancement option that is the great unknown—an option that has all but been ignored in the glee of search. What exactly are these companies doing with your keyword data? It's naive and idealistic to think that your anonymity and privacy are preserved, as they claim. Three or four searches are more than enough to identify you personally. How many people are interested in Chihuahuas, swimming, and surrealist art, and check online movie timetables for your town or city? Just you, I'm afraid.
Building "paper dolls" of people from their search activities is risky business for the people under scrutiny. In the 1980s, when databases became popular, there was an outcry against cross-referencing for the purposes of constructing such paper dolls. One crystallizing moment was that of the naked man shot to death by Special Forces when answering the doorbell of his apartment in Berlin. Authorities had put together building services records from multiple sources and concluded that his apartment must house a Cold War terrorist cell. Tension was so high when the door was opened—due to a paper doll expectation—that he was shot even though he held no weapon and had nowhere to conceal one. He merely maintained the apartment as a place to meet his lover, and so the paper doll interpretation was entirely wrong.
Subsequent restrictions on database cross-referencing seem to have almost vanished. My city council had no trouble acquiring my address from the local motor registration board when I forgot to re-fill a parking meter. I must be a parking meter terrorist. Obviously, connecting bits of data from here and there is a highly political matter: Sometimes there's a persuasive argument for it, and sometimes there isn't.
A search engine company is unlikely to break down someone's door, but search is clearly an important piece of social infrastructure and easily politicized. It's infrastructure that informs average people, gives voice to minorities, and acts as a democratizing force—if left in its raw form. Like the courts or the government, it's subject to corruption. As it stands, none of us have any idea what search companies do with the data that they collect. We have no idea how they order the results that they display for us, and certainly no one is saying what information search companies should be allowed to push to their users. A few advertising codes apply, but nothing beyond that.