Click here for contact details, other articles and personal details

The future of surveillance

October 2001

I’m writing this a week after the attack on the World Trade Centre, when security is very much in people’s minds. Already, some people are predictably demanding that strong encryption technology should be forced to have a back door so that governments can intercept messages if necessary, and civil liberties groups are equally predictably crying fowl. The sad fact is that terrorists and criminals already have encryption technology and won’t be deterred from using it just because it is made illegal, so it is perhaps too late for such actions. What is needed instead is a means of collecting communications prior to encryption. Although terrorists can hide anywhere, it will soon be possible in principle to put monitors almost everywhere in a country, and while such blanket surveillance would be expensive, it would be a fraction of the price of conventional warfare and probably a lot more effective.

Imagine ‘smart snow’, where millions or billions of tiny monitors that cost a few pence each to make could be scattered over a wide area, deployed as a mist from high altitude aircraft. Each could monitor activity in a very small area and communicate interesting activity to the distant observer via ad-hoc networking. They would cooperate with adjacent sensors devices using self organisation techniques to form sophisticated surveillance networks. The decreasing physical size of electronics could soon bring useful devices down to a few millimetres or less, so that it would be very hard for a group of would-be terrorists to ensure that none are present. However much an area is inspected, there might always be one nearby that has worked its way into a crack. In a warfare situation, they may even be able to discreetly assist missiles to home in on them if need be.

Of course, countries are very large. Afghanistan is 650000 square kilometres for example, and a sensor may only be able to cover a region of 10 square metres effectively after allowing for wastage. 65 billion sensors would be needed to cover the entire country. If they could be produced for 10 pence each, the cost of policing this whole country would be £6.5 billion, a fraction of the cost of significant military action. Because they could overhear conversations, the only way people could be certain of privacy under such a surveillance system would be to use deep caves or carefully cleansed buildings, and to carefully ensure that no discussion is every undertaken outside. Cleansing an area could be made very difficult, since the devices might only transmit their findings on certain conditions, and may then do so in a very short burst, making them hard to locate. They would not need to transmit voice, just a few bits to indicate that a suspicious conversation has taken place.

With a few minor enhancements, even caves and buildings would not be immune. Coating some chips with sugar would encourage ants to carry them to nests, which often would be near buildings. Air currents could be used by suspending chips like dandelion seeds. Even aerogels may be used, filled with helium to make them as light as air, so that they would float around. Chips could be stuck on insects that naturally search out human habitation for food. So surveillance on a grand scale could be technologically feasible.

Other novel means of surveillance using insect-like robots have already been developed to a point. In due course, the main threat of information theft or destruction might not come from Code Red virus descendants, but from insect robots that climb into your PC, or even the keyboard, physically connect to the circuitry, intercept or modify the data, and then travel home. We will need much more than encryption to address such problems.

Meanwhile, back home and today, police forces are being forced to ensure that speed cameras are painted in bright colours. In spite of this, the UK has more surveillance than any other country so true anonymity and complete privacy are already memories. I don’t like being monitored continuously by cameras when I go into town, or when I browse the web, but such is everyday life already and the surveillance level is tolerable, even if I don’t like it much. What worries me a little more is that as sensor technology improves, and as devices of all kinds become are interconnected, a very detailed picture of my life is increasingly available to government, police, and hackers. While I may not be a criminal, I am concerned that much of this data will be backed up more or less indefinitely. Come the revolution, a hostile government sifting at lightning computer speeds through this data could very quickly identify the people it considers as undesirable, and perhaps I might be on their target list. And it may not just be the authorities. Perhaps as cameras become cheaper we will see them everywhere, even in children’s toys, allowing a small girl to see the world through Barbie’s eyes. But older, less innocent children, might use them to spy on their teachers. In the adult world, coupling cheap cameras to number-plate recognition and face recognition enable communities to police themselves, monitoring the behaviour of every visitor and car. They could export blacklists onto the web of those that they consider to be a problem. You could find your photo on a list of suspects on a web page simply by being in an area at a time when a crime occurred. Official surveillance might be the least of our problems.

Such a future world could feel very oppressive. Most of us aren’t criminals, but still occasionally break the letter of the law, and frequently do things that some group in society would disapprove of. While we may want some surveillance to control crime rates, we will need protection against abuses of surveillance too. We may have to more carefully regulate the uses to which new technologies may be put, and to limit the distribution of data collected by surveillance.

But there are some new dangers that come not from people but machines. A machine can have no intent to either obey or disobey the law. However, machines can already evolve software, using various techniques based on genetic algorithms. While trying to implement some wholesome software, machines could occasionally accidentally write and implement code that would defraud people or organisations. Part of its ongoing optimisation may be to replicate this software, and to build an evolution environment in which it can constantly adapt to the marketplace. Of course, it may use encryption to ensure that no-one interferes with its integrity. As the level of artificial intelligence increases, many such pieces of software, from different sources, might interact and perhaps co-operate, even interbreed in the evolution environment. Such a scenario could result in many pieces of code that have evolved and interbred on a word wide platform, their origins or indeed distant ancestry, and thus ownership very tenuous indeed. Without any original malicious intent, we could have very efficient programs running on the net, heavily encrypted, adaptive, evolving, distributed and self replicating, but with complete disregard for the law. Many variants of this scenario are possible, so our security and privacy might be threatened by the very tools we use. Other developments such as peer to peer computing offer the ability for algorithms to be distributed around the net so that they cannot easily be located and destroyed, making policing of these potential problems much more difficult.

Given the recently increased threats to our security from terrorist groups, we can certainly expect a good deal more surveillance in one form or another. Balancing the use of such technology against our need for privacy and against potential misuse, or its use by hostile powers, will become increasingly difficult.