This is not a call for some halcyon day of some spring of previous years. Nor, is this an abysmal call for the lower technology world so many profess to want unless it means giving up their BlackBerry. This is supposed to be a thoughtful discussion of what a world looks like when privacy is not a given and is instead an opt-out form of tyranny. This is a discussion of why privacy is a national security issue in America.
The three basic security services are called confidentiality, integrity, and availability. Some strange people who don’t have the ability to think flexibly talk about authentication (mixture of integrity and availability) and non-repudiation (mixture of confidentiality and integrity). We won’t blame those people. Though most attacks against information will be a hybrid of all three confidentiality is what makes the news most often (called a breach). When the personal identifiable information (PII) of people is compromised we refer to breaches or hacking. The release of the PII isn’t the actual exploit rather it is the outcome.
Confidentiality is the principle we look towards for privacy. In the legal realm the Supreme Court of the United States (including the current Chief Justice Roberts) has referred to the penumbra of privacy as a right of people. Even the National Security Agency was charged with the role of protecting people’s privacy as far back as the 1976 by then President Ford. The right has weakened substantially as criticism of the penumbra has been slowly eroded. More importantly what has eroded is in a day of technology is the willfulness to object to breaches by government and corporations.
There are a lot of definitions for national security. In my opinion national security is the principle of nation-state sovereignty (self rule) and ability to resist hostile or eminent danger from domestic or foreign adversaries. I would predicate that national security requires a certain amount of secrecy exists, and at the same time too much secrecy could be bad for society in general. If my construction of the national security argument is not too flawed it is important to national security that privacy exist as a principle for the people.
Information can come at an organization sideways. Time magazine in a 1990 article “And bomb the anchovies” stated that before large military operations the local pizza deliveries to the Pentagon would escalate. The local companies could simply look at the sales numbers and know if the United States was about to go to war. Whether hyperbole or not the pattern of one set of information being used to diagnose a separate behavior is a form of information leakage. Unanswered is what is the Pizza company’s particular responsibility to restrict the information of who is purchasing pizzas? If you answered none you would be right.
Now enter the world of big data. When we look at the concepts of information aggregation allowed for by large unstructured data sets that can now be manipulated information leakage would be egregious. Consider in the Pizza example you wouldn’t just need to call up the Pizza delivery owner you could tie into a just in time inventory system and collect the data in real time. Perhaps you could tie into a GPS based “Where is your Pizza now” system and simply track how many stops are made at the Pentagon. The use of cyberspace as an attack platform against privacy of people to get to national security would look something like this (using the pizza example” for simplicity.
- Query: Is America getting ready for war, and who is planning it?
- Pizza deliveries increase by 20% in year to year comparison on a particular day. Data sets come from Pizza delivery situational awareness (groan), point of sale clearances, and just in time inventory controls.
- Pizza order is paid for my senior colonels who are tied to plans and operations at the Pentagon (information is from credit card used online, point of sale system, and harvested biography information).
- Two of said colonels have been listed as working in section dealing with likely target country (information harvested from on line biography, press releases and FaceBook status updates)
- The two colonels are friended on FaceBook with X, Y, Z staffers and work for Assistant Deputy Secretary so-and-so (information available through FaceBook and various sharing services used by junior staffers)
- Social network analysis of those entities show that they in previous engagements were at such-and-such level of coordination and planning (historical data)
- In previous events time between Pizza delivery and operation commencement was a window of 72 to 96 hours (historical data)
- At “potential target a” American citizens credit cards report a significant uptick in departures (not that Visa controls wouldn’t pick that up already).
- So on and so on…
If the data set is large enough and can be manipulated quickly enough you don’t need to violate “national security systems” to violate national security. Some authors have discussed this as the inference problem. Unfortunately most of the information needed in the pizza example can be purchased in real time from a variety of data warehousing agencies. This same form of data intrusion could be accomplished against a corporate entity. If the entity has the ability to track searches by geo-location like a Google, Yahoo, or Bing then the data set becomes even richer. What are they interested in and what is the area that they are looking at in the way of knowledge. You don’t ‘google’ something you know. You can do analysis on what is searched on as much as what is not searched on.
By using the failure of privacy protections for citizens an adversary could use data aggregation to bypass significant efforts of national security systems. This though is not how government tends to play this issue. The pro forma argument is that we must give up a certain amount of privacy for national security. The other side of this is that you can’t have national security without privacy. This isn’t about using FaceBook or Google it is about the ability to manipulate and infer information based on the data rich sets available.
Since much of the data is available for purchase as “anonymized” there is some feeling of protection from untoward remote surveillance. This is in error. As an example use the colonels above. How many colonels are in the United States Military? How many work at the Pentagon? How many would be working late and “willing” to feed their workers pizza? How many becomes a set that can be easily deanonymized.
I used one specific example that isn’t to sensitive and worthy of discussion. There is a whiff of fear uncertainty and doubt contained in this little discussion, but spending some time looking at the data warehousing stake holders quickly takes the stink off. The one thing that is obvious is that privacy is inherently a requirement for national security. Go ahead and don’t believe me. Believe President Gerald Ford.