What could be more revealing than a list of one's search queries? The efficiency of finding what we need on the Web encourages us to quest away--whether we're researching a car purchase, puzzling out some medical symptoms, wondering what happened to an old friend or (gasp) groping for erotica. "Your search record involves aspirations and dreams," says Marc Rotenberg of the Electronic Privacy Information Center. "It becomes almost a reflection of what's in one's head." And, as we learned recently, when America On-line temporarily released the search history of thousands of customers for the use of researchers, those reflections can be retained by search companies--and, ultimately, exposed.
In the AOL case, the records were supposedly anonymous, but since people commonly type in their own names and addresses into search engines, it's often trivial to identify who is searching. Indeed, The New York Times was able to deduce the identity of a 62-year-old widow in Georgia who researched Italian vacations, termites and hand tremors.
The intimacy of our searches has led Rotenberg and other privacy experts to urge companies like Google, Yahoo and Microsoft not to retain such logs. And a bill proposed by Rep. Edward Markey (Democrat of Massachusetts) would set limits on the length of time such data could be stored. But the top researchers in the search field argue that such limits would ultimately be destructive. These wizards believe that the information extracted from studying the way individuals search has been crucial in raising the quality of search to its present level. "Our searches have improved dramatically because we have that data," says Alan Eustace, Google's senior vice president of engineering and research. Furthermore, they contend that without the information, they would be severely hobbled in further improving their products. "If you don't have such data, there would be significant compromise of the user experience in the future," says Prabhakar Raghavan, Yahoo's head of research.
In particular, these companies hope to introduce schemes by which one's searching success would be enhanced by previous behavior. For instance, when a fishing buff types in "bass," he or she would get a different result than the same search from a Paul McCartney fan.
As you would expect, search companies like Google and Yahoo say that protecting the information on those search logs is assigned highest priority. (Though I bet if you asked folks at AOL about this before the semipublic release of its users' most intimate data, they would have said the same thing.) Yahoo even has a security group referred to informally as "the paranoids," whose motto is "We worry about these things so you won't have to."
But even if the companies are flawless in protecting that information, there's still reason to worry. The federal government has already expressed interest in such records, and if the data are subpoenaed the companies must turn it over. It's also not much of a stretch to envision a scenario (i.e., another terrorist attack) that would pressure the companies to submit to a sweeping request for search logs, just to see if there's a sleeper cell or two out there. (In the process, why not investigate those people looking for nude pictures, or swapping Madonna tunes?)
It is possible to search without a trail. So-called anonymizing services can mask your identity when you surf. And some smaller search engines do not keep session logs. The big players, though, are betting that you'll stick around--because, they say, access to your dreams makes their searches great.