Wikimedia v. NSA: Another Court Blinds Itself to Mass NSA Surveillance
We all know justice is blind. But that is supposed to mean that everyone before it is treated equally, not that the justice system must close its eyes and refuse to look at important legal issues facing Americans. Yet the government continues to convince courts that they cannot consider the constitutionality of its behavior in national security cases and, last week, in an important case for anyone who has ever used Wikipedia, another judge agreed with that position.
A federal district judge in Maryland dismissed Wikimedia v. NSA, a case challenging the legality of the NSA’s “upstream” surveillance—mass surveillance of Internet communications as they flow through the Internet backbone. The case was brought by our friends at the ACLU on behalf of nine plaintiffs, including human rights organizations, members of the media, and the Wikimedia Foundation.1 We filed a brief in the case, too, in support of Wikimedia and the other plaintiffs.
The judge dismissed the case based on a legal principle called standing. Standing is supposed to ensure, among other things, that the party bringing the lawsuit has suffered a concrete harm, caused by the party being sued, and that the court can resolve the harm with a favorable ruling.
But the U.S. government has taken this doctrine, which was intended to limit the cases federal courts hear to actual live controversies, and turned it into a perverse shell game in surveillance cases—essentially arguing that because aspects of the surveillance program are secret, plaintiffs cannot prove that their communications were actually, in fact, intercepted and surveilled. And without that proof, the government argues, there’s no standing, because plaintiffs can’t show that they’ve suffered harm. Sadly, like several other courts before it, the judge agreed to this shell game and decided that it couldn’t decide whether the constitutional rights of Wikimedia and the other plaintiffs were violated.
This game is mighty familiar to us at EFF, but that doesn’t make it any less troubling. In our system, the courts have a fundamental obligation to conclusively determine the legality of government action that affects individuals’ constitutional rights. For years now, plaintiffs have tried to get the courts to simply issue a ruling on the merits of NSA surveillance programs. And for years, the government has successfully persuaded the courts to rely on standing and related doctrines to avoid doing so.
That is essentially what happened here. The court labeled as “speculative” Wikimedia’s claim that, at a minimum, even one of its approximately one trillion Internet communications had been swept up in the NSA’s upstream surveillance program. Remember, this is a program that, by the government’s own admission, involves the searching and scanning of vast amounts of Internet traffic at key Internet junctures on the Internet’s backbone. Yet in court’s view, Wikimedia’s allegations describing upstream—based on concrete facts, taken from government documents— coupled with a plaintiff that engages in a large volume of internet communications were not enough to state a “plausible” claim that Wikimedia had been surveilled.
On the way to reaching that conclusion, and putting on its blindfold, the court made a number of mistakes.
The Government’s Automated Eyes Are Still Government Eyes
First, it appears the court fundamentally misunderstood Wikimedia’s claim about upstream surveillance and, in particular, “about surveillance.” As Wikimedia alleged, “about surveillance” (a specific aspect of upstream surveillance that searches the content of communications for references to particular email addresses or other identifiers) amounts to “the digital analogue of having a government agent open every piece of mail that comes through the post to determine whether it mentions a particular word or phrase.” The court held, however, that this type of “about” surveillance was “targeted insofar as it makes use of only those communications that contain information matching the tasked selectors,” like email addresses. But what the government "makes use of" is entirely beside the point—it is the scanning of the communications for the tasked selectors in the first place that is the problem. To put it into a different context, the government conducts a search when it enters into your house and starts rifling through your files—not just when it finds something it wants to keep. The government's ultimate decision to “make use of” the communications it finds interesting is irrelevant. It is the search of the communications that matters.
Back of the Envelope Gymnastics
Another troubling aspect of the court’s decision was its attack on the probabilities Wikimedia assigned to the likelihood of its communications being intercepted. Given that Wikimedia engages in a large volume of Internet communications, Wikimedia alleged that—even assuming a .00000001% chance that any one particular communication is intercepted—it would still have a 99.9999999999% of having one of its communications intercepted. The statistic was used to illustrate that, even assuming very low probabilities for interception, there was still a near-certainty that Wikipedia’s traffic was collected. But the court attacked Wikimedia’s simple statistical analysis (and the attack tracked, to a great degree, arguments made in the government’s declarations that the court purportedly did not consider). The court seemed to believe it had seized upon a great flaw in Wikimedia’s case by observing that, if the probability of any given communication being intercepted were decreased 100% or 1000%, the probability of one of Wikimedia’s communications being intercepted would similarly drop. The “mathematical gymnastics” the court believed it had unearthed were nothing more than Wikimedia using an intentionally small (and admittedly arbitrary) probability to illustrate the high likelihood that its communications had been swept up. But even if the court disagreed with the probabilities Wikimedia relied on, it’s not at all clear why that would justify dismissing the case at the outset. If it turned out, after development of the record, that the probabilities were off, then dismissal might be appropriate. But the court cut the case off before Wikimedia had the opportunity to introduce evidence or other facts that might support the probability they assigned.
Someone Else Probably Has Standing, Right?
Perhaps most troubling was the court’s mistaken belief that the legality of upstream surveillance could be challenged in other ways, beyond civil cases like Wikimedia or our ongoing case, Jewel v. NSA. The court asserted its decision would not insulate upstream from judicial review, which—according to the court—could still receive judicial scrutiny through (1) review from the Foreign Intelligence Surveillance Court (FISC), (2) a challenge by a criminal defendant, or (3) a challenge from an electronic service provider. None of these options is truly a viable alternative, however. First, the FISC (until very recently) did not have adversarial proceedings—it only heard from the government, and its proceedings remain both far more limited and more secretive than a regular court’s. Second, a challenge from a criminal defendant won’t work either, because, to date, the government has explicitly refused to disclose—even where defendants are notified of the use of FISA surveillance—whether their communications were obtained using upstream surveillance. And, finally, in the nearly 15 years (or more) the government has conducted upstream surveillance, we’re not aware of any service provider that has challenged the legality of the practice. Indeed, given that upstream is done with the cooperation of telecoms like AT&T and Verizon—the same telcos that did not challenge the NSA’s bulk collection of Americans’ call records for over a decade—we're not holding our breath for a challenge anytime soon.
Instead, we need the courts to tackle these cases. Upstream surveillance presents unique constitutional issues that no federal court has seriously addressed. It's time the federal courts stepped up to the challenge.
On Wednesday, we asked the Ninth Circuit Court of Appeals to hear Jewel v. NSA—our long-running case challenging the NSA’s upstream surveillance. Our clients are a class of AT&T customers, and we have direct evidence of a facility at AT&T’s offices in San Francisco (and some evidence of several others elsewhere) where the government seizes and searches traffic at the Internet backbone. Nevertheless, 7 years into the case, we still have not received a decision on the merits from a court. The arguments the government has made in our case are similar to those in Wikimedia, although the specific issue before the Ninth Circuit on Wednesday was a different government attack – one urging even more delay.
The NSA has been engaging in wholesale surveillance of the Internet backbone in the US for nearly 15 years now, surveillance that has admittedly affected millions of Americans. It’s long past time for the courts to take off the blindfold and consider whether that surveillance violates our constitutional rights.
- 1. The plaintiffs in the case are: Wikimedia Foundation, the National Association of Criminal Defense Lawyers, Amnesty International, Human Rights Watch, PEN American Center, Global Fund for Women, The Nation Magazine, the Rutherford Institute, and the Washington Office on Latin America. For simplicity’s sake, we’re just going to refer to the group of plaintiffs as Wikimedia.