The Google Subpoena Case: A Google Victory

You may also like...

17 Responses

  1. Dissent -there’s too much assuming of the conclusion. Part of it has to do with the fact that the reasoning is obscure. But that doesn’t make it sinister or pretextural. That part of the ruling is interesting:

    “Additionally, this is not a case where the Government does not have the benefit of any information with which to form some basic methodology –the Government has already been to the pond and fished, so to speak, with data from AOL, Yahoo, and Microsoft, and it would not have been unreasonable at this stage to have required the Government to assist the Court in its determination of relevance by providing the Court with more information on its plans for the information sought from Google”

    That is, if it was a fishing expedition, THEY’VE FISHED ALREADY!!! In multiple ponds. So that would seem to be ruled out.

    I think the Govt. just doesn’t want to tip its hand as to its legal strategy.

  2. David Zaring says:

    Uh, not to be critical of my generous hosts, but although the government did narrow its initial search request, as one does in these sorts of cases, I note that its narrow search request was granted. Frankly, as it should have been – we’re past the days of fishing expeditions claims in discovery- that’s what protective orders are for.

  3. David writes: “Uh, not to be critical of my generous hosts, but although the government did narrow its initial search request, as one does in these sorts of cases, I note that its narrow search request was granted.”

    No, that’s not quite correct. As I said in my post, there were two prongs to the government’s request: (1) 50,000 URLs and (2) 5000 search queries. Discovery for the URLs was granted but NOT discovery of the queries. The request for the queries was the part of the government’s request that raised privacy concerns and was the focus of most of the media attention when the government first made the request (in much broader form). And with regard to the 5000 search queries request, part of the government’s argument was that it would be protected by a protective order, but the court still didn’t grant the discovery request.

  4. But the Court didn’t grant the request because they called it “duplicative”, which seems a much smaller victory than one might get the impression from secondary sources:

    “From this Court’s interpretation of the Government’s general statements of purpose for the information requested, both the sample of URLs and the set of search queries are aimed at providing a list of URLs which will be categorized and run through the filtering software in an effort to determine the effectiveness of filtering software as to certain categories. Both sources of the URL “test set” list seem to be open to the same sorts of criticism by Plaintiffs in the underlying litigation. The content of these objections are not germane to the Court’s determination of whether the information sought is relevant under the broad dictates of Rule 26, but the actual similarity of the two categories of information sought in their presumed utility to the Government’s study indicates that it would be unreasonably cumulative and duplicative to compel Google to hand over both sets of proprietary information. To borrow the Government’s vivid analogy, in order to aid the Government in its study of the entire elephant, the Court may burden a non-party to require production of a picture of the elephant’s tail, but it is within this Court’s discretion to not require a non-party to produce another picture of the same tail. Faced with duplicative discovery, and with the Government not expressing a preference as to which source of the test set of URLs it prefers, this Court exercises its discretion pursuant to Rule 26(b)(2) and determines that the marginal burden of loss of trust by Google’s users based on Google’s disclosure of its users’ search queries to the Government outweighs the duplicative disclosure’s likely benefit to the Government’s study.”

  5. Seth — the search queries were the part of the request that posed the greatest concern for privacy. In this regard, there is a difference between the search queries and the URLs. For the purposes of the government’s study, there may not have been much difference, but from the standpoint of privacy, there is a significant one.

    I basically read the opinion as saying politely that the government’s request is unnecessary, but the court will throw the government a tiny bone and deny the demand for information that can implicate people’s privacy. Even before the opinion, the government had already practically backed down — from requiring two months worth of searches (millions, perhaps billions of searches) to requiring just 5000. I see this as the government saying, in a sense: “Please, please just give us something — we’re now only asking for just a miniscule fraction of what we wanted.” And even this small scaled-back request was still denied. But the court is being somewhat nice about it — it provides the government the 50,000 URLs, which do not strike me as all that contentious or even all that important — even though the court has all but held that the government really hasn’t made a good justification for the need. In other words, the court seems to be throwing the government a bone so it can save a tiny bit of face.

    But in the big picture, the court seems to be suggesting that this is as far as the government can reach, and it isn’t very far. The government had originally asked for all the URLs in Google’s database (as of July 2005) and all search queries in a 2-month span. The extent to which the government backed down is extraordinary. And then it still lost partially in the case. So comparing what the government had wanted to what it ultimately got, the government did not do well at all. And I read the court’s opinion to be only very reluctantly giving the government it’s bone — had the government not backed down and scaled back its requests, I bet that there’s a chance that the government might have lost entirely.

  6. Daniel – Interestingly, I see the decision almost entirely the reverse. I see the court as basically deferring to the government, but tossing *Google* a tiny bone on the issue of the *perception* of privacy, so that *Google* can save face.

    My view is that the first request for all the data was basically out of ignorance. They’re lawyers (no offense meant!) they had no idea it was so absurdly large. Previously, they’d gone down to one million URLs and “a random sampling of one million search queries submitted to on a given day”. Then they got into some posturing and went up on the demands. But from day one, this has only been about a statistics study, never, ever about particular users’ behavior. They’ve never wanted to invade anybody’s privacy, and have all sort of protective orders for that purpose. Note in fact the government is now insulated from certain objections to the study, because they can say that they tried to do better in terms of sampling, but other constraints required them to modify the methodology, so they’ve produced the best evidence legally possible.

    “But in the big picture, the court seems to be suggesting that this is as far as the government can reach, and it isn’t very far.”

    Again, I view the court as saying the exact opposite:

    “Nothing in this Order is intended to indicate how the Court would rule on the original broad subpoena or on any follow-up subpoena. The Court’s decision on this Motion to Compel reflects the limited use to which the Government intends to put the information produced in response to the subpoena. In particular, this Order does not address the Plaintiffs’ concern articulated at the hearing about the appropriateness of the Government’s use of the Court’s subpoena power to gather and collect information about what individuals search for over the Internet.”

    I view that as saying the government might be able to get more if they made a better argument, and that the judge is specifically *NOT* making an overall privacy ruling, especially when combined with the “duplicative” reasoning.

    Thus regarding: “So comparing what the government had wanted to what it ultimately got, the government did not do well at all.”

    I very strongly don’t think the original request is a valid basis for comparison. It’s like the civil lawsuits where the original demand is a zillion dollars in damages – it’s just for show, and each side knows it’s almost certainly not going to be the ultimate outcome (and here, I think the government didn’t even know they really didn’t want it, and couldn’t handle it if they got it!).

  7. Well, Seth, I guess we’ll just have to disagree on this one. I read the court’s caveats over the reach of its decision as basically saying: “It’s a balancing, so this opinion is not a statement that privacy always wins.” So yes, one cannot use this opinion in cases where the government’s justification is different, because the test is a balancing one, and the court’s decision is based in part on the weakness of the government’s side of the balance.

    I don’t find the analogy to civil damage demands to be apt. This does not strike me as a case of aiming high to improve one’s negotiating posture. I don’t think that the government’s initial demands were just for show.

    And in the end, the government didn’t really get much in this case. I find it very hard to interpret the result as a victory for the government.

  8. I think the initial demands were more out of ignorance, but it’s *like* the civil damages posturing over amounts. That is, a comparison that they asked for 4? 8? billion URLs, got only 50,000 URLs, is like so-and-so asked for 4? 8? billion dollars in damages, got $50,000 dollars in damages. Pushing this metaphor, one could analogize the user searches to requests for punitive damages. I’d argue the proof that the big demand was just their starting point, is that even before this went to court, the government was willing to go down to one million URLs on the database, and down on the searches. It seems a lot more sensible to me that the first big request was because they didn’t really understand it, and maybe figured they could go down (as they did) much more easily than go up.

    The victory I see for the government is that the government was able to get Google to contribute to its study, with giving almost no justification. Remember, *critically*, the big war is about the Child Online Protection Act (COPA) law, NOT, NOT, NOT, about search engine user privacy (though I’d definitely count the outcome as a *PR* victory for Google).

  9. David Zaring says:

    Sorry – that’s correct about the queries, but I don’t think the government isn’t considering the ruling a victory. Full disclosure: I used to work on this case for said government. But I’m with Seth on this one, despite the nice skeptical language in the opinion, the result was consistent with that turned over by the other search engines.

    But to make a larger point, privacy is a very flexible term to use to avoid discovery – look how much useful information it swallows up in the discovery-like FOIA. To tweak my very brilliant host some more: would it even be a good idea to have third party privacy protected in discovery?

  10. David Zaring says:

    That is, protecting third party privacy as a general matter in cases like this one, and through anything other than a protective order (which are controversial in their own right, see Arthur Miller on that, but strike me as less problematic than the denial of discovery of material that’s going to tell us something of interest about the internet).

  11. David — I can’t quite tell where your descriptive analysis ends and your normative analysis begins. Descriptively, the government got a ton less than it originally had wanted, even under the very permissive standards for subpoenas. The key part of the battle was the search queries, not the URLs, and the number of search queries turned over was zero.

    Regarding the normative issues, I do find third party privacy to be very important and worth protecting, especially when the entity seeking the information is the government. I disagree with Dionisio, the Supreme Court case holding that subpoenas are not searches under the Fourth Amendment. I also disagree with Miller, Smith, and the other third party doctrine cases. I won’t rehash my arguments here, but you can read them in my book, The Digital Person, or in this paper.

    What I find interesting about this case is that the standards for allowing discovery of information pursuant to a subpoena are remarkably permissive (far too permissive in my view) — and yet the court still denies the government’s search query request even after the government has whittled down its request to nearly nothing.

  12. David Zaring says:

    Fair enough – descriptively we disagree. Normatively we might as well. I’m not sure how relevant it would be to the constitutionality of COPA, but, as someone interested in the internet, I’d be interested in seeing the results of the study that these subpoenas will permit to happen. Let’s hope that I – and everyone with similar interests – gets to see the results.

  13. David,

    I’d be very interested in an explanation for why the government needed the information. If the rationale was to test filtering software, why couldn’t it have found URLs on its own and created its own search queries? I’m really straining to understand why the government even bothered to make the requests in the first place. And why did the government want such a large amount of information initially and then back down? This undercuts its credibility to some extent in its argument that it needs the information.

    In other words, why can’t the study be conducted without the information from Google? The government’s request makes little sense to me.

    Anyway, the reason why descriptively I believe that the case was a victory is because subpoenas are rarely quashed, and the subpoena standards are very pro-disclosure.

  14. David writes: “I’m not sure how relevant it would be to the constitutionality of COPA, but, as someone interested in the internet, I’d be interested in seeing the results of the study that these subpoenas will permit to happen. Let’s hope that I – and everyone with similar interests – gets to see the results.”

    That seems to me to be a problematic justification to use the subpoena power to gather information — to conduct a study that would be of interest to people. I’d sure like the power to commandeer Microsoft’s software team to make software I want, or to commandeer another company’s information to conduct empirical research for me. I can dream up a lot more cool research projects and studies . . . I wanna go work for the DOJ so I can find some way to force companies to help me out on them!

  15. Regarding: “The key part of the battle was the search queries”

    Actually, I’m not sure that’s true for the *government’s* *goals* *in* *COPA*. That is, it’s pretty interesting that they gave almost no *reason* why they needed it apart from sampling. It would be amazingly anti-climactic if it turned out that the search queries are overkill.

  16. On second thought, I should back off from that. It would be better to say there’s two parts to what the goverment is trying to study, call them “existence” and “popularity”. And “popularity” (which needs the search queries) is probably the more important part of the two.

    I don’t think the judge really grasped what the study is trying to do, which is understandable given that the government was so vague, which is also understandable given that they don’t want to lay out their strategy.

    I suppose if it’s really important, we may see another round of this case later on.

  17. David Zaring says:

    Okay, I can’t resist one more comment. I agree with Seth’s comments above about the existence and popularity being a possible result of the study. Dan and I don’t completely disagree on the oddness of broad subpoena powers, but I think it could be less bad than requirements that subpoenas be targeted and their purpose be directly relevant to whatever (though that would certainly be good for the lawyers). And I don’t think the government is comandeering anything here – it’s analyzing information that Google wanted to keep private. Just because the government could have performed other kinds of analysis doesn’t mean that it shouldn’t have been allowed to perform this kind of analysis. And with that, I cede the floor.