JQP note: This article re-appears here by kind permission of the editors (one of whom is the Head Research Otter) over at PSUK, for whom it was originally written. I urge anyone concerned about the balance between citizen autonomy and state control to read and contribute to their project.
Sometimes, I wonder whether the human habit of gerontocracy can survive in a high-bandwidth society. Clearly, the experienced are better administrators than those who haven’t learned the hard way; not necessarily better policy makers, but certainly better administrators. But when Moore’s Law has taken over your world and the pace of change, particularly in the sphere of data generation, acquisition and analysis, has become so fast that the map can shift under the territory faster than you can draw, how can we ask people born before the fax machine to understand and legislate about technology?
“This consultation covers an important topic that affects us all. The capability to use communications data to protect the public is being eroded by new technology. In seeking to maintain that capability, the Government must strike the right balance between public safety and privacy.” — UKGov
Surveillance has historically been about desperately trying to get enough data to analyse, so that you can figure out what just happened. Then people got into the whole computer thing, and it became about trying to get enough data to analyse that you could figure out what’s happening right now. That’s really hard, and well beyond the ambitions of any government or state intelligence agency (I include Scotland Yard in that definition) prior to 1975. Except GCHQ but they were always a bit special.
When the PC revolution hit, though; when we started being able to collect data faster and more comprehensively than ever before, the agencies started to run into the data/information paradox. Data is 1s and 0s. Data is irrelevant without analysis; and if your mission is to chart and track the world and all it’s myriad, wonderfully chaotic humans then data is history unless the analysis is timely. Real-time, one might argue. And the problem they’ve had ever since is that data acquisition and storage capacity far outweighs either speed of analysis, or accuracy of it. Unfortunately, they’re getting pretty good at data analysis from their own point of view; they have systems (Bayesian mathematics is useful here) which can determine a ‘typical profile’ and then apply it to data and tell you who doesn’t fit. Marketroids love that kind of thing; they only need to be accurate 1-5% of the time to make a profit.
However, as Cory Doctrow eloquently pointed out in Little Brother, that kind of accuracy rating sucks when you’re talking about arresting people. An aside on this book; it does what Cryptonomicon and Snow Crash between them tried to do, but does it better and has the added bonus of being (for the most part) accurate to the modern internet in terms of its details. The protagonist m1k3y lays out the problem with false postives in clear, layman’s terms; if you want to find something rare (e.g. terrorists ) you need an incredibly accurate search algorithm. The data analysis techniques in use by law enforcement, surveillance organisations and even ISPs are nowhere near accurate enough, or fast enough, to deal with the flow of data in real time.
So let’s return to this consultation. The quote above can be unpacked quite easily; what it means is the government know that their ability to spy on the populace is being compromised by the pace of change. Their attitudes can’t keep up; people just keep having new ideas, and the data flows just keep going up. What they want to store is traffic data; and if you have no problem with people profiling who you talk to, what you read and what MMORPGs you visit, then you need to ask yourself this: why should they care? About you specifically, I mean? The answer they will give you is “We won’t know until we’ve looked.”
This is industrial-era thinking. This is the world before the Wars trying to set policy for the Internet Generation. They do not understand the nature of privacy as it applies to high-bandwidth civilisation because they do not understand either the technology or the international, independent culture which it feeds. They have no way of comprehending, let alone implementing, “the right balance between public safety and privacy” because they understand neither the threats nor the concept of personal autonomy. To them, every citizen is a segment in their hive and the citizen’s data belongs, not to the citizen, but to the state.
The government now know that they can’t use comprehensive spying techniques to monitor their population in real time. Their dream of TIA is now beyond their intellectual capacity; they can’t win an arms race when the kids know the tech better than they do. Therefore, they are trying to pass a law which permits them to out-source the effort and the data storage, so that they can make use of one of our ideas: distributed processing networks do data analysis better than monolithic hierarchies. They want the ISPs to solve their surveillance problems for them.
Write to them. Get informed by reading the consultation PDF and surrounding material (have a look for data on the ISP industry’s fight against the RIPA: El Reg is a good start). Figure out what you think, and then let this guy know:
Nigel Burrowes Communications Data Consultation Room P.5.37 Home Office 2 Marsham Street London SW1P 4DF
You can email them direct at communicationsconsultation at homeoffice dot gsi dot gov dot uk.
Get informed. Get involved. And never trust anyone over 25.
 Estimates by vaguely creditable people seem to vary between 2000 potential Islamist militants in or with access to the UK, and 4000 from the Torygraph. I can’t imagine them offering conservative figures on this topic, so let’s take that as the top plausible limit. That’s 4000/60,975,000 = 0.000065601 chance in 1 of finding a terrorist. That’s a pretty accurate search algorithm you want there. And that’s without considering noise interference in the data, badly designed search paramaters due to racist law enforcement, the fact that not all Islamist militants are ever going to be actual terrorists, and any number of other real-world factors that really screw with your math.