"When you are looking at that blank search box, you should remember that what you fill can come back to haunt you unless you take precautions."-- Pam Dixon, Executive Director, World Privacy Forum
The many millions of Americans who routinely seek information on myriad matters through Internet search engines like Google naturally assume (as the search providers assure them) that the subject matter of their inquiries will remain private. In the summer of 2005, however, that expectation was challenged by a Justice Department demand filed in federal court, seeking extensive access to Google search data. Despite government assurances that the results would be used only in aggregate form, and that the demand was not seeking information which could “personally identify” the individuals who made the search requests, Google strenuously resisted this unprecedented demand.
The source of the government’s interest in such data has nothing to do with national security. It stems instead from litigation challenging the constitutionality of the Child Online Protection Act (COPA), which Congress passed in 1998 after the Supreme Court invalidated the “indecency” provisions of the Communications Decency Act. COPA targets sexually explicit material that could be harmful to minors, and imposes criminal penalties on those who post such material for commercial purposes. On its second trip to the Supreme Court in 2004, COPA fared badly because Internet filters seemed to offer a less restrictive alternative to the tough criminal penalties imposed by the statute. The Justices, however, left open the possibility that the government might still convince them that less drastic approaches (such as filters) would not adequately protect young Internet users from harmful material.
In the quest for such evidence, the Justice Department demanded extensive digital data from several search engine companies. The government explained that access to such files would help to “understand the behavior of current Web users, [and] estimate how often Web users encounter harmful-to-minors material in the course of their searches.” Three other subpoena targets (AOL, Yahoo, and MSN) complied with the government’s demand; Yahoo, for example, insisted “this is not a privacy issue” because of the aggregate nature of the intended data use. But Google – the dominant search engine – insisted it could not surrender such potentially sensitive data without severely compromising the privacy of its subscribers and inevitably chilling their use of the Internet to obtain vital information. The Justice Department rejected Google’s privacy plea and initially requested the contents of 5,000 randomly selected search requests. Google argued that giving up this information could cause a loss of its traffic if some Web users refused to use the search engine for fear that the government would be allowed to review their search requests. Following Google’s protest, the government scaled back that initial request and in mid March, 2006, a federal judge ordered Google to comply with a subpoena seeking a listing of 50,000 randomly selected URLs in Google’s database. The judge’s order, however, expressed concern about the privacy interests of Google’s users that could be implicated if the government had not scaled back its initial request and had actually sought a log of search queries.
The quest for data that will sustain the government’s defense of COPA is surely appropriate, since the Supreme Court essentially invited such a justification. But to build that case at the expense of the long and confidently assumed confidentiality of Internet search engine users seems an excessive cost, not only to privacy but also to freedom of expression in cyberspace given the potential chilling effect.