I wrote a column this past July on the NSA's PRISM database and the government surveillance apparatus that motivated it. You may recall that one central theme of my column was that while the five PowerPoint slides leaked by Edward Snowden and initially published by the Washington Post and Guardian newspapers were pretty innocuous, the overall government surveillance apparatus that has been building for the past 40 years was far from innocuous. We have since learned that Snowden had much more to offer the media that was exceedingly provocative (cf. http://www.washingtonpost.com/world/national-security/black-budget-summary-details-us-spy-networks-successes-failures-and-objectives/2013/08/29/7e57bb78-10ab-11e3-8cdd-bcdc09410972_story.html?wpisrc=al_excl ), but that's the subject of another column.
The five slides extensively reproduced by the media remained classified after the leaks and subsequent reproduction in the media. I included a screenshot one of these slides in my July column. Though it appeared in the printed version of Computer, it was removed from the IEEE digital library version. Pull up a chair and let me tell you a story about how our surveillance state can control what you see in your professional publications.
Spillage is government-speak for information that ends up where it shouldn't. The formal definition in the Committee on National Security Systems Glossary of April, 2010 is a “Security incident that results in the transfer of classified or CUI information onto an information system not accredited for it (i.e., authorized) for the appropriate security level…. [e.g.] …whenever classified data is spilled either on an unclassified information system of to an information system with a lower level of classification.” ( https://www.cnss.gov/Assets/pdf/cnssi_4009.pdf )
Edward Snowden's five PRISM PowerPoint slides were examples of classified information spillage. Intuitively, after most of the world has seen something, it isn't secret anymore in any meaningful sense of the word, so journalists and broadcasters treat the spillage as public information. Spillage, and the whistleblowers and leakers that provide it, is a necessary byproduct of investigative journalism. Spillage has always been a core component of journalists' “ground truth data.”
However, Snowden's leaks were so embarrassing to the U.S. Government, that the DoD Defense Security Service sent out an official Notice on June 11, 2013 reminder-warning contractors to avoid spillage on their networks.
“Contractors shall not, while accessing the web on Contractor's unclassified systems, access or download documents that are known or suspected to contain classified information. Classified information, whether or not already posted on public websites, disclosed to the media, or otherwise in the public domain remains classified and must be treated as such until such time it is declassified by an appropriate U.S. government authority. It is the responsibility of every Contractor to protect classified information and to follow established procedures for accessing classified information only through authorized means.
“Contractors who inadvertently discover potentially classified information in the public domain shall report its existence immediately to their Facility Security Officers (FSO). Companies are instructed to delete the offending material by holding down the SHIFT key while pressing the DELETE key for Windows-based systems and clearing of the internet browser cache. Subsequently, administrative inquires and adverse reports are not required. These procedures apply only to the inadvertent exposure to classified information in the public domain.” (http://www.dss.mil/documents/isp/Contractor_NOTICE_posting.pdf )
Why would the government do this? Certainly not to stem the flow of Snowden’s spillage – those horses were out of the proverbial gate. The government wants federal contractors (e.g. Booz Allen Hamilton) to understand by whose largesse they owed their economic fortunes. This was more than a gratuitous act - this notice put the contractors on notice that they had better ramp up the policing of their IT infrastructure, or else.
What did this demand for network hygiene accomplish? Intimidation, pure and simple! The DSS Notice provides the Government additional leverage against contractors that don't aggressively police their workforce for potential whistleblowers and leakers. The last sentence in the Notice exposes the charade of using spillage as the trigger of this additional scrutiny: the notice only applies to “public domain” information - i.e., yesterday's news. By the time the DSS posted this notice, newspaper copies of the original slides had already passed through contractors offices, break rooms, and waste baskets, and no doubt prompted lively conversations in cafeterias and around water coolers. But those activities don't get audited (at least not yet)! To be compliant with this DSS Notice subcontractors had to report spillage to the FSO – and that meant creating audit trails that the government can inspect. The DSS Notice is simply an Orwellian tactic to deal with thought crimes – the step before a visit to the Ministry of Love.
Similar signals were sent to the media who reported the leaks. The Monterey Herald first reported an Army-wide blockade of the UK Guardian newspaper's website to achieve a “vigilant command climate” in DoD-speak on June 26, 2013 (http://www.montereyherald.com/local/ci_23546947/guardian-news-website-blocked-at-presidio-monterey ). The Guardian was the original source of Snowden's leaks, so the DoD took careful aim by blocking access from DoD computer facilities to the newspaper's website http://www.theguardian.com/world/2013/jun/28/us-army-blocks-guardian-website-access . Lt. Col Damien Pickart confirmed that this also applies to all “websites that re-report information first published by the Guardian” ( http://www.usnews.com/news/blogs/washington-whispers/2013/06/28/blackout-defense-department-blocks-all-articles-about-nsa-leaks-from-millions-of-computers ), so the digital blockade was pervasive. Of course, this behavior is not new: the DoD does the same for WikiLeaks and presumably for any other news source that provides access to embarrassing stories. According to Pickart, one of the primary rationales for the blackout is economic – server hygiene is costly so it's preferable to simply block access. Think about this for a while. Ask yourself this question: Is spillage on unclassified networks the real core of the DoD's cybersecurity problems? Spillage is not even low hanging fruit - from the perspective of risk, it's discarded biomass.
As intimidating as these DoD shots across the bow were, they pale in comparison to the British government's reaction to the Snowden leaks. They actually raided the Guardian's offices. You see, the Brits lack a first amendment and apparently have much more latitude when it comes to imposing prior restraint on free speech than the US. They demanded the hard drives that contained Snowden's materials from the Guardian ( http://www.theguardian.com/world/2013/aug/20/nsa-david-miranda-guardian-hard-drives ). Rather than turn over the hard drives, Guardian editor Alan Rusbridger chose to destroy them. According to Rusbridger, two GCHQ security experts witnessed the physical destruction. “Whitehall was satisfied, but it felt like a peculiarly pointless piece of symbolism that understood nothing about the digital age.” Reuters reported that the request to hand over or destroy the hard drives came directly from British Prime Minister David Cameron ( http://www.reuters.com/article/2013/08/21/us-usa-security-snowden-britain-idUSBRE97K0G920130821 ). The choice, as the Guardian saw it, was to comply or risk the British government's closure of the newspaper.
Technology service companies were included in the wave of government intimidation as well. This past August, Lavabit abruptly shut down its operation after the FBI obtained a search warrant for metadata (a so-called pen register) for a specific account. It has been reported that the account holder of interest was Edward Snowden who used the account to advertise press conferences he held in the Moscow airport. Lavabit refused, was threatened with criminal contempt, appealed, received a search warrant for “all information necessary to decrypt communication sent to or from all Lavabit email accounts including encryption keys and SSL keys.” (http://www.wired.com/threatlevel/2013/10/lavabit_unsealed/ ) Understanding that an anonymous email service that gives up authentication keys is for all intents and purposes out of business anyway, Lavabit owner Ladar Levison simply closed the doors rather than comply with the order (cf. ( http://techcrunch.com/2013/10/03/lavabit-founder-details-government-surveillance-of-secure-email-while-documents-disclose-epic-trolling-of-feds/ ). Levison remains under a gag order. It remains to be seen how far the government pushes the contempt case against him. In reaction to the Lavabit closure, another email anonymizing service, Silent Mail, preemptively followed suit ( http://silentcircle.wordpress.com/2013/08/09/to-our-customers/ ).
It appears that the newest target of government wrath might be academic freedom. In early September, a professor of computer science at Johns Hopkins was instructed by his dean to remove a blog post critical of the NSA from the university's mirror site. Why the dean did this is unclear at this writing (cf. http://www.theatlanticwire.com/politics/2013/09/johns-hopkins-university-falls-victim-nsa-chilling-effect/69219/ ; http://www.propublica.org/article/johns-hopkins-and-the-case-of-the-missing-nsa-blog-post ), but it's likely that he was externally motivated, as faculty blog oversight is not normally within an purview of an academic dean. The offending blog post is reproduced at http://arstechnica.com/security/2013/09/crypto-prof-asked-to-remove-nsa-related-blog-post/
So that's the backdrop against which the rest of this story must be placed. I submitted my column on July 14, unaware of the latest DoD DSS missive three days before and the subsequent implications that would have for at least one IEEE corporate subscriber to the digital library. An attentive facility security officer of Solers, Inc, a beltway government contractor, sent a spillage security alert to employees concerning the PRISM screenshot that appeared in my column (in particular in the digital library version of my column to which IEEE members had access). The IEEE received a copy of this alert.
This security alert took on a life of its own as it ricocheted through the Computer Society's and IEEE's editorial offices, eventually landing in the office of the IEEE General Counsel and Chief Compliance Officer, prompting a statement to Editors of IEEE Transactions, Journals, Magazines, Newsletters and publications officers that included these three paragraphs:
“--- IEEE is a U.S. nonprofit corporation and, therefore, is considered a U.S. person subject to U.S. laws, including those dealing with the possession or publication of classified or other national security information. “
“--- U.S. laws do not treat such classified or other national security information as being “in the public domain” just because they become available through the Internet or other media. Instead, they remain classified or otherwise protected materials and government property in spite of that availability.”
“--- Legal precedent in the United States has held that the government may not prevent the publication of leaked classified or otherwise restricted government, which is protected under the First Amendment of the U.S. Constitution relating to the freedom of the press. However, after such publication the government may in certain cases still prosecute the publisher for possession or publication of such materials as criminal violations of applicable U.S. law.”
The legal precedents include New York Times v. United States and the subsequent prosecution of Daniel Ellsberg and Anthony Russo under the Espionage Act of 1917. The Supreme Court held in Times v. U.S. in 1971 that the U.S. government failed to satisfy the burden of proof that is required for a prior restraint injunction discussed above, and that the Times was free to continue to publish the Pentagon Papers, but (and this is a critical conjunction) that the government was free to prosecute Ellsberg and Russo after the fact. As it turned out, the resulting Ellsberg and Russo prosecution resulted in a mistrial because of the misconduct of the Nixon administration's prosecution (White House Plumbers operations, etc.). Hence, Ellsberg and Russo were not acquitted, there was no definitive Supreme Court ruling, and therefore nothing added to the body of case law.
In this way an image that had already appeared in virtually every news outlet was removed from the electronic copy of my column in the IEEE digital library, along with all references thereto. How did we get to the point where one of our cherished professional societies became intimidated by the government in this way? The answer is to be found somewhere in the intersection of uncertain case law, DoD digital blackouts of media, intimidation of government contractors, and pressure on journalists and authors who might be critical of the government and the surveillance state.
Computer authors receive feedback continuously – but generally not from their publisher's attorneys. In my case, I received a call from the IEEE Legal and Compliance Department on July 18 expressing concern over the unsettled state of case law surrounding spillage vis-à-vis my July column. Sympathetic to the IEEE's concern, I recommended, without hesitation, that the IEEE simply remove the offending image, leaving behind the text and caption as is, and substituting for the image something like the spillage alert or a URL to the image available on Wikipedia. In so doing the IEEE would simultaneously ameliorate their legal concerns while remaining on the right side of history.
In my opinion, the removal of all textual references to the image as if to pretend that the image never appeared in the first place will be judged poorly by history. However, this is an area over which intelligent people may disagree. Not-for-profit professional societies are not the best perches from which to launch first amendment test cases. Distinguished not-for-profit media organizations such as NPR ( http://www.indiewire.com/article/outrage_review_spiked_for_naming_names and PBS ( http://www.policymic.com/articles/43793/citizen-koch-pbs-kills-koch-brothers-critical-documentary-for-fear-of-offending-them ) have been pressured to pull controversial content occasionally, and have done so without permanent damage to their reputations. Although the motivations may have been different (protect privacy (NPR) vs economic pressure (PBS) vs. threat of litigation (IEEE)), not-for-profit organizations are of necessity risk averse. In fact, commercial television is not immune – e.g., ABC's pulling of the “Brill's Content” reporting. First Amendment zealots will wish that professional media organizations have complete editorial license over what they decide to release, but wishing won't make it so.
Although I don't think for a moment that the government would ever prosecute an academic professional society on a spillage charge, the IEEE cannot survive if large numbers of government employees and contractors cancel their memberships. That's where the intimidation factor comes in, and one of the reasons that the spillage notice was posted in the first place. From a historical perspective, the more dangerous threat isn't spillage of classified information, but the spillage of government intimidation! No publisher, media outlet, professional society, NGO, corporation or individual is immune from this. The government will go to any length to maintain its appearance of control.
By the way, my response to all parties involved was to encourage all not-for-profit professional societies, and scholarly publishing companies, to raise the issue of how they might all stand together to encourage legislative reform in the area of classified information that's inadvertently divulged into the public space (in other words, spillage). The current atmosphere where the government is unwilling to declassify such information — while it simultaneously increas es the risk to government subcontractors, publishers, and media outlets (such as with the DSS notice of 11 June 2013) —is , is hard to reconcile with the need to advance scholarly and scientific inquiry. This problem can only get worse as the information needs increase in such critical areas as digital security and privacy, genetics, and cloud control, not to mention thorny Constitutional issues. I would hope that through combined lobbying efforts , effective and meaningful change might take place.
Henry David Thoreau said that “Our inventions are wont to be pretty toys which distract our attention from serious things. They are but improved means to an unimproved end, an end which it was already but too easy to arrive at.” (http://thoreau.library.ucsb.edu/thoreau_life.html ).
History has shown that the velocity of innovation usually exceeds our ability to manage it for the public good. Such is the case with the government's use of digital surveillance. It's not that we cannot in principle manage it, but rather that we're intellectually and politically ill-prepared to manage it. Technology evolves without check until abuses begin to alarm literate and informed segments of the population.
In his own way, Snowden was calling attention to the fact that it is far easier to create and deploy surveillance technology than to responsibly use it. Western society has always had a problem with technology stewardship, often deferring to unbridled technology change for its own sake. Such is the advance of weaponry, pesticides, the exploration and use of fossil fuels and nuclear energy, misuse of pharmaceuticals, non-FDA approved medical compounding etc. Phrases like the Cutter incident, the thalidomide crisis, Love Canal, Bhopal, Chernobyl, Fukushima Daiichi, Deepwater Horizon, Exxon Valdez, and the Johnstown Flood effortlessly slide into our vocabulary as silent witness to our technological immaturity.
The IEEE’s excising of innocuous spillage from its digital library was understandable for a professional society that relies on member dues and subscriptions for revenue and lacks the resources for lengthy court cases.
Stanford law professor Lawrence Lessig is well known for his poignant observation that software code may actually provide more regulation over our behavior in cyberspace than the law ( http://codev2.cc/download+remix/Lessig-Codev2.pdf ). The Snowden leaks confirm what can happen when a government concurrently creates laws while also developing computing systems specifically designed to circumvent these very laws it is tasked to enforce. In such a case code and law are not two dimensions of cyber-regulation, but adversaries, one to the other. This introduces a unique spin on cyber-dystopia: where the government's code acts as its agent whenever Constitutional protections become too burdensome. This isn't regulation by code, but oppression by code. In Lessig's terms, government regulates the code directly so as to better regulate behavior indirectly: code is power, and government code is absolute power. We don't have to look overseas to see how a government can use internet technology against its citizens.