I believe that the record of personal interests and actions which is captured in the logs of search engines and other internet sites is just as revealing, private, and intimate as the personal data that resides on our hard disks, in the filing cabinets of our homes and offices, or hidden on papers slipped under our mattresses... Neither the government nor anyone else should have any more right to access our personal data via Google than they would have to access the same data stored in our own private places.
The goverment demands via subpoena that Google turn over records of one week of private users' searches so that it can build a case to support a question of policy embodied in a law (COPA) that has been rejected by the courts. Although Google is not a party to any proceedings in this non-criminal matter, the government claims that Google still has an obligation to provide the personal information since Google as "[t]he non-party witness is subject to the same scope of discovery under [Rule 45] as that person would be as a party to whom a request is addressed pursuant to Rule 34. [Fed. R. Civ. P. 45, advisory committee's notes to 1991 amendment.]" See Page 5 lines 23 to 25 of the Government's Motion to Compel.
I believe the government seeks Google's cooperation since it will be less burdensome to get the data from Google than from us individually. It can not be argued that it would be easier to get the aggregated data concerning the search behavior of millions of americans from Google rather than issuing a blanket subpoena forcing all americans (as non-party witnesses) to deliver up a record of their personal searches. Given the current state of computer technology, we undoubtedly have means to ensure that even if individual americans were forced to give up their data, it would be possible to build systems that would remove just as much "personally identifiable" information from what was submitted as Google is capable of removing. The government could easily, although not cheaply, build a system that takes the individual submissions from individual private citizens, aggregates them, and processes them to produce an effective duplicate of the data that Google might submit in response to this subpoena. The government is essentially demanding that Google build and employ precisely such a system to satisfy the government's demands -- if Google hasn't already done so for other purposes.
Given that the results of the two possible subpoena's would be substantially identical, except in cost of implementation, we should not support a right to one subpoena unless we would also support (at least in principle) a right to the other. If we can accept the government's right to subpoena all of us in this case, then there may be a case for seeking to reduce the burden of obtaining the data by using Google as a convenient and cost-effective proxy for individual service.
Before we address the question of how burdensome it may be to respond to a subpoena, we must first answer whether or not the government has the right to the data which the subpoena demands. I believe no such right exists in this case. If the government's demands were less broad, or involved a specific prosecution rather than an issue of policy, it is entirely possible that I would feel otherwise... In this case, Google should be commended for refusing to respond to the government's subpoena. Any other services that did provide the demanded information should be condemned and shamed for having done so.