Googlethink

I type the letter p into Google’s search box, and a list of 10 suggested keywords, starting with pandora and concluding with people magazine, appears just beneath my cursor. I type an r after the p, and the list refreshes itself. Now it begins with priceline and ends with pregnancy calculator. I add an o. The list updates again, going from prom dresses to proxy sites.

Google is reading my mind—or trying to. Drawing on the terabytes of data it collects on people’s search queries, it predicts, with each letter I type, what I’m most likely to be looking for. The company formally introduced the automatic recommendation of search terms in 2008, after a few years of testing. It’s been tweaking the service, which it calls Google Suggest, ever since. This past spring, it rolled out the latest enhancement, which tailors suggestions to a searcher’s particular city.

Google Suggest, like the similar services offered by other search engines, streamlines the discovery of information. When you click on a suggestion, you arrive at a page of search results, and the accompanying advertisements, a little faster than you would have, had you typed out the query yourself. Google Suggest is technically remarkable. It testifies to the power of cloud computing—the serving-up of software and information from big, distant data centers rather than from a computer’s own hard drive. When I typed that first p, the letter was beamed across the Internet to a Google server in a building hundreds of miles away. The server read the letter, gathered 10 popular search terms beginning with p, and shot the list back to my screen. This intricate data processing took less than a second. It felt magical.

It felt a little creepy, too. Every time Google presents me with search terms customized to what I’m typing, it reminds me that the company monitors my every move. The privacy risks inherent in such long-distance exchanges became apparent in February, when three European researchers revealed that they had used intercepts of some Google Suggest traffic to reconstruct people’s searches. Alerted to the breach, Google quickly added a new layer of security to the transmissions, but the researchers claim that vulnerabilities remain.

I like Google—it’s a cuddly company, and endlessly helpful—but I also resent it. It’s like a nosy mother, intent on knowing everything her children are doing and thinking. Worse, it’s like a meddlesome mother, the kind who can’t let her kids do anything on their own. Start typing a keyword, and she immediately butts in, trying to finish it for you. At first you enjoy the hyperactive solicitousness. But then you begin to bridle. You’re being smothered.

Matthew Crawford, in his book, Shop Class as Soulcraft, writes eloquently about our modern affliction of “displaced agency.” As corporations rush to anticipate our every need and preference, serving up a small selection of market-tested options the instant we require them, we are left with little room to act for ourselves. “We have too few occasions to do anything,” Crawford writes, “because of a certain predetermination of things from afar.” Everything becomes easier, but less satisfying.

Software programmers are taking the displacement of personal agency to a new level. Relentlessly focused on making their programs more “user friendly,” they’re scripting the intimate processes of intellectual inquiry and even social attachment. We follow their scripts when we click on one of Google’s keyword suggestions, and we follow them when we select from a list of categories to describe ourselves and our relationships on Facebook. These choices are convenient, but they’re not our own. They’re generalizations masquerading as personalizations.

Eric Schmidt, Google’s chief executive, once remarked that he looked forward to the day when Google would be able to tell him “what [he] should be typing,” which, if I’m interpreting the statement correctly, also means that Google would be telling him what he should be thinking. Such a service, Schmidt said, would be the product he’s “always wanted to build.” I can’t say I’m looking forward to it.