This story appeared on Network World at

http://www.networkworld.com/columnists/2006/013006bradner.html

'Net Insider

King George I on privacy

By Scott Bradner, Network World, 01/30/06

At first the news of the U.S. government asking Google for a week's worth of everyone's Internet searches was mostly astonishing in how unexpectedly expected it was. Since the Internet has been made the bogeyman for just about every ill that troubles mankind, why not get a head start by finding out in advance what everyone is looking for?

The details behind the headlines, when they came out, mostly added confusion. This case may turn out to be far less than it first appeared, but it could easily lead to an attempt at a fundamental realignment between the rights of the individual and the government.

In an attempt to rescue a 'Net censorship law that the Supreme Court had tossed out, the government used a subpoena to ask Google and the other search engine companies for a whole pile of information about how people search the Web and what they find.

The original request to Google was for "[all] URLs that are available to be located on your company's search engine as of July 31, 2005" and for Google to "produce an electronic file containing [all] queries entered into the Google engine between July 1 and July 31 inclusive."

All in all, a huge amount of information. The subpoena said that the government wanted just the search strings "without any additional information that would identify the person who entered any individual search string." Google said no; AOL, Microsoft and Yahoo all said yes to similar requests.

What is the problem with these requests, considering that the government is specifically asking that identifying information be removed? Leaving aside the fact that many searches do reveal personal information about the searcher, it's very hard to figure out just what the government is trying to prove with an analysis of this information.

The law that was overturned required that Web sites with naughty material on them use a form of positive identification to be sure that a child was not trying to access them. The court said local filters were a way to get to the same place while interfering less with the constitutionally protected rights of Internet users who were not children. The analysis of the search terms could - not shockingly - show that people look for naughty things using Google. (Note that a local filter could block such searches.) An analysis of URLs might show that some of them contain naughty words. That too is hardly a shock and also something that a local filter can block.

I'm left with the nagging feeling that the case is just an excuse to find out if there are characteristics of search strings that can be used to ferret out bad guys of one sort or another. If the government thinks that turns out to be the case, how long will it be before the government "asks" the search companies to become its agents and turn over additional information that would identify the person who entered specific search strings in the future? In fact, if the analysis of this set of data finds suspicious searches, do you really expect the government not to demand information on who did the search?

It is a natural thing for governments to think that the privacy of the individual is an impediment to security. They thought that in the time of King George I, and little has changed.

Disclaimer: Harvard has been a witness to this tendency since long before the first King George, but the above is my own nagging feeling.