• Choose a category
• All
• Classic-Fiction

By Root 1706 0

getting things wrong. In early 2010, when Google launched Google Buzz, its Twitter-like service, they did not take appropriate care in protecting the identities of many of their users, disclosing their contact lists in the erroneous belief that no one would mind such intrusions into their privacy (even Andrew McLaughlin, Google’s former senior executive and the deputy chief technology officer in the Obama administration, was trapped in the Buzz trap, as many of his former Google colleagues appeared in his contact list). Though Google executives downplayed the significance of the accident by claiming that no one got seriously hurt in the debacle, in truth we don’t know how many new names and connections were added to the KGB’s databases as a result. The real costs of Google’s misjudgment cannot be immediately calculated.

Think, Search, Cough

Every time we post a greeting to our friend’s Facebook wall, Google the name of our favorite celebrity, or leave a disapproving comment on the website of our favorite newspaper, we leave a public trail somewhere on the Internet. Many of these trails, like the comment on the newspaper’s site, are visible to everyone. Some, like our Google searches, are only visible to us (and, of course, Google). Most, like that odd comment on the Facebook wall, fall somewhere in between.

Fortunately, we are not alone on the Internet—at least one billion other users are also blogging, Googling, Facebooking, and tweeting—and most of our information is simply lost in the endless ocean of digital ephemera produced by others. This is what privacy scholars call “security by obscurity.” In most cases, obscurity still works, even though there are more and more exceptions to this rule. Ask anyone who has difficulty finding a job or renting an apartment because something embarrassing about him or her appears in Google searches or on Facebook. Nevertheless, aggregating these tiny digital trails into one big data set—sometimes across entire populations—could produce illuminating insights into human behavior, point to new trends, and help predict public reaction to particular political or social developments. Marketing and advertising companies understood the power of information a long time ago. The more they know about demographics, consumer habits, and preferences of particular customer types, the more they can tailor their product offerings, and the more they can make in sales as a result.

The digital world is no different. The history of our Internet search says more about our information habits than our patron files in the local library. The ability to identify and glean “intent” from a mere Internet search, matching advertisers with customers looking for their offerings, has allowed Google to turn the advertising business on its head. Thus, in addition to running the world’s most successful advertising agency, Google also runs the most powerful marketing intelligence firm. This is because Google knows how to relate Internet searches to demographics and other searching and purchasing decisions of its customers (e.g., what percent of New Yorkers who searched for “digital camera” in the past twelve months ended up searching for “deals on iPhones”).

But we’re not just looking for better iPods and new deals on plasma TVs. We are also seeking information about people and places in the news (“has Michael Jackson died?”), about broader cultural trends (“what are the best novels of the decade?”), and, of course, about solving problems—mostly trivial but some important—that constantly pop up in our lives (“how to repair a broken washing machine”).

There are many seasonal variations to how often we search for particular items (searches for “stuffed turkey” predictably increase before Thanksgiving), but the frequency of queries for most items is usually fairly consistent. Thus, whenever there is a sudden spike in the number of Google queries for a given term, it probably indicates that something extraordinary has just happened; the likelihood is even higher if the search spike is limited to a particular geographic