As widely expected, Google notified a federal court on Friday that it will not comply with a subpoena by the U.S. Department of Justice to produce a random sample of the searches performed on its search engine in 2005.
However, few observers expected the company to put that refusal into an indignant 25-page brief that depicts Justice Department lawyers as burdensome, poorly thought out and technologically naive.
The Justice Department wants to use abstract data from Google and other top search engines to build a picture of the amount of pornography on the Web and how effective today’s software filters are at blocking it. The aim is to reinstate the 1998 Child Online Protection Act. That law imposed fines and jail time for Web operators who provided adult content that was then viewed by minors.
The law was judged unconstitutional by a federal court in Philadelphia, in a decision upheld by the Supreme Court in 2004. But the high court said that case might be reopened if technological changes made filters less effective. The government has said it must formulate a picture of the adult content on the Web in order to test the effectiveness of filters. Yahoo!, AOL and MSN Search have all complied with amended government requests for data scrubbed of personal identification details.
But Google’s response to the subpoena points out that the techniques it uses to search and index the Web are so unique that prosecutors would not be able to construct a picture of what material is actually out there on the Internet without having access to Google’s proprietary technology.
Similarly, search users who customize their searches or who use an available Google filter to screen out adult-content results will also skew the data, the brief said.
Google also pointed out that because of continual changes to its search algorithm, the results produced by queries on a keyword can differ greatly over time. “In no meaningful way can it be said that a past week’s worth of search queries will yield URL responses as performed by ‘present day users of the Internet,’”, according to the Google filing.
The search engine was particularly harsh in its critique of the government’s aim to use the requested search data to understand the search behavior of current Web users. “This statement is so uninformed as to be nonsensical,” it said in the brief, pointing out that without personal identifying data, raw Google search queries will combine requests from adults and minors, and even humans and automated robots.
In fact, Google said in its brief, it has actually begun to receive pornography requests that it believes are in reaction to the government subpoena. One individual has even developed a software application that sends a random porn query to Google along with any other search request the user may make—helping to undermine the government’s purpose in subpoenaing the Google data.
Google maintained in its filing that the government request would tie up employee resources and processing power, possibility causing slowdowns for Web search users. It characterized as “inadequate and unrealistic” the government’s offer of expert help in producing random samplings of both keyword searches and Web URLs.
The company also made the point that complying with the DOJ request would have a chilling effect on the trust Google users place in its search. “If users believe that the text of their search queries into Google’s search engine may become public knowledge, it only logically follows that they will be less likely to use the service,” the company stated.
The Justice Department has until Feb. 24 to respond to Google’s filing. A U.S. District Court judge was originally slated to hear oral arguments from both sides on Feb. 27; but that date has been pushed back to March 13.
The American Civil Liberties Union filed its own brief on Friday urging that the court reject the DOJ’s Google request as intrusive and burdensome. “The government is not entitled to go on a fishing expedition through millions of Google searches any time it wants, just because it claims that it needs the information,” ACLU staff attorney Aden Fine said in a statement.
The ACLU is the central plaintiff in the lawsuit challenging the constitutionality of the COPA measure. In its response to the California court, the organization said that if the Justice Department wins the release of the Google search data, the ACLU will file its own subpoena for Google data to assess the reliability of the random sampling. This request would encompass the total number of Web pages in Google’s database, the frequency of index updates, the configuration of its server network and “how Google’s search engine functions and produces results based on the input of queries”—all highly proprietary information that Google has never released, for fear it will be abused by Web optimizers or copied by competitors.
The case is one of two currently shining a spotlight on the ethical dimensions of the search industry. Google has been in the odd position of winning plaudits from privacy watchdogs for its stand against releasing search data to the U.S. government, while facing criticism from human-rights advocates and the U.S. Congress for allegedly bowing to wishes of the Chinese government and making some dissident Web sites hard for Chinese users to find.
