Ever since I saw the call records story breaking this morning I've been sitting here with a very visceral visual of exactly what Big Brother knows about me.
See, what I do for a living is, I study gene networks, and specifically methods for monitoring what changes in a network of genetic and biochemical interactions when a perturbation is introduced. We use a lot of graph theory and network analysis in this field. I'm a data miner by trade, although my focus is mostly on cleaning up the signal so others can mine it more effectively, because of course, in networks, errors add up and false connections are made and you get the wrong answer, which means (in biology) that you waste time pursuing irrelevant leads.
So when I think of what a database of all of our call histories looks like I get the heebie jeebies. Because I know with a fair degree of certainty what kind of connections they can look for in that data, and what kinds of events they can flag. With these records, they know who you call, they know when you call them (and can maintain a history that stretches over years) and they know exactly where you call them from.
- Big Brother knows everyone that you care about. Not just your family of birth but your family of choice.
- Big Brother knows who all your friends are, and who their friends are, and friends of their friends. He knows when you make a new friend, and he knows when you stop talking to someone. He knows when that friend you haven't heard from in years suddenly turns up in your life. Big Brother knows that after you got that call from that old friend, your social network lit up with calls back and forth to other old friends to exchange hot gossip.
- Big Brother knows when you're making personal cell phone calls from work. He knows which way you drive home, if you're making calls from your car on the way.
- Big Brother knows that you know a guy who knows a guy. Say, for instance, you call a friend a lot and that friend calls a guy who gets busted for dealing drugs. Big Brother knows that you called your friend and then your friend called his dealer and that this happened to happen in that sequence several Friday nights in a row.
- Big Brother knows when you're calling your illicit lover to arrange an assignation, and he knows when you call your wife from the hotel room to tell her you're working late.
- Big Brother knows when you call that guy you were dating who blew you off and just as quickly hang up the phone.
- Big Brother knows when you went on a trip, where you went, who you called while you were away and when you got back. Big Brother also knows who takes care of your pets while you're gone.
I'm sure there's many more examples of patterns that we could define. But the point is: it's absolutely unprecedented in history for a government to have that kind of detailed knowledge of its citizens' connections and the patterns of their private lives. It brings a whole new level of 'total' to the word 'totalitarian'.
With the start and end time of a phone call, the identities of the originator and receiver (easily mined from other less-secure sources even if the telcos strip them off what they sell to the government), and the cell phone tower location, literally any pattern of associations among people, businesses, places and times can be mined out if there's a smart geek at the data mining end who knows how to formulate the pattern as a query. And your node may not have identifying information attached to it, but it may be associated through a pattern with a node that does -- someone who has somehow gotten on the radar, say, by being rounded up at an anti-war rally or registered to vote as a Democrat or busted for holding a small amount of pot.
The reason that this thing has got me so disturbed today is that I am a smart geek, and I could formulate that query and direct a programmer to implement it, much as I direct students to implement queries in a genomic database. I have the imagination, and the training in looking for patterns in data, to understand how such information could and would be used. In the world of database queries, if you can dream it (formulate the criteria) you can do it.
Sure, the power to know such things seems innocuous. But if you wouldn't trust the Chinese government under Mao with information about who the friend you love to talk to most actually is, then you shouldn't trust your own government. The power to use those connections just needs to be animated by a small amount of will to use them wrongfully and, as I read somewhere else today, you could be stripped of your rights as an American faster than you can say Padilla.
We need to find a way to bring home to non-geeks what the power of such a database can be, to take it out of the realm of geekiness and show how easily it could be used to attach consequences to seemingly innocuous actions and patterns. People should be madder than hell about this. It should drive George Bush's approval rating into the negative numbers.
For medical data, to use that data for research, we have to meet elaborate standards under HIPAA to strip any personally identifying information off the data. What we need is a HIPAA for communications and consumer data. Force the corporations to file off the serial numbers on the information they track about our lives. Let the corporations mine our purchasing and resource use patterns for trends all they want, but let them not connect those trends to individuals unless a warrant is issued to connect the abstract to the individual.