DREAMS NIGHTMARES OF THE DIGITAL AGE

DREAMS & NIGHTMARES

OF THE DIGITAL AGE

Recently a lot of worrying has been done about the Internet. People who are concerned about anarchy tend to see it as a thrumming hive of villainy, where crazy extremists exchange nerve-gas recipes whereas people who are fearful about government see the Internet as the perfect tool for Big Brother.

If the Internet had gone wrong in just the right way, either of these threats might have resulted. But it hasn't. When you want information on the Internet, either you know exactly where it is or you don't. If you already know where it is, then the fact that it happens to be on the Internet is irrelevant; you could just as well get it by fax or mail. If you don't know where it is, you have to use a piece of commercial software called a search engine. You type in a few key words or phrases, and the engine tries to find World Wide Web documents that contain them. When you try this, the search engines will turn up either irrelevant texts (documents in which the words nerve, gas and recipe happened to appear separately) or references to the general idea of nerve-gas recipes.

The Internet is the only place where worrying about something makes it less likely to happen. Every time someone posts a message to a newsgroup about the danger of nerve-gas recipes on the Internet or a journalist writes an article about the same topic, which is then available on the Internet, a little more noise is added to the system, and any would-be terrorist who tries to search for a nerve-gas recipe finds literally hundreds of thousands of red herrings.

Even if a person was able to download a string of words billed as a nerve-gas recipe, he would face a long time of anxiety if he got the ingredients together and sat down to mix them. Who would be crazy enough to trust an anonymous recipe that, if it goes wrong, will kill him? The very anonymity that makes it possible to post nerve-gas recipes makes the people who have posted them untraceable and unaccountable and hence difficult to trust.

Crypto (cryptographic technology on the Net) is going to change that, however. If we are going to worry about a technology, let's worry about crypto.

Crypto is a collection of protocols. 'Protocol' here means a set of rules governing an exchange of data, agreed to by all participants. The number of core protocols in crypto is fairly small.
One example is the digital signature. Data can be given a digital signature that, like an ink signature on a contract, verifies that they were signed by a particular person. It is stronger than an ink signature because it proves absolutely that the data have not been meddled with. Another protocol is public-key encryption, which, to make a long story short, can encode documents so securely that no government on earth can decode them in a reasonable amount of time. These and a few other basic protocols can be combined to make more complicated ones, such as digital cash--a system for transferring real money (not just credit-card numbers) from one person to another, anonymously and untraceably.

An advanced protocol that has been getting a lot of attention from high-technology developers has to do with reputations. Anyone who has found his E-mail box filled with junk mail or done a Web search that turned up hundreds of thousands of irrelevant documents may have wished that his computer would select only the stuff he finds interesting. This reduces to a problem of assigning an unforgettable reputation--an index of reliability, tailored to the user's personal interests and biases--to the source of each piece of information. This in turn is related to digital signatures.

Put it all together, and in a few years we might have something to worry about. Someone searching for a nerve-gas recipe might be able to tell his search engine, 'Show me only the good stuff,' where 'good' means that which is highly regarded by fellow terrorists. The recipe would arrive encrypted so securely that no government-watchdog agency could read it. The digital signature would prove that it came from a well-reputed chemist and that its contents had not been tampered with on the way. The terrorist could mix up some amount of the stuff, leave some of it sealed in a container in a public place and then send an untraceable E-mail to the authorities telling them where to find it and demanding that a certain ransom be paid lest more be dumped into a subway vent during rush hour. The payment would then be made, not through an exchange of cash but through an untraceable, digital-cash transfer to an anonymous electronic bank account--with no risk to the terrorist.

There is almost no limit to the ways in which generalized cryptographic technology could be used by bad guys. Wiretapping will be abolished as crypto telephones become widespread. The battle to control electronic transmission of child pornography will soon be lost. It will be much more difficult to track down drug dealers by following their money trail.

All this explains why the U.S. government has long been worried about crypto and has tried to combat its spread with increasingly foolish measures, such as declaring cryptographic software to be a munition. From a careers perspective, a crypto programmer with a non-U.S. passport is a good thing to be.
Nearly all Silicon Valley companies must build some form of crypto into their products to make them competitive. Since U.S. regulations forbid the export of such technology, these companies simply order that portion of the work be done overseas.

Naturally we look to the government (for all its faults) to exercise some control in these situations. But one thing that makes crypto unique is its potential to cripple government. Governments can't function without revenue, which they get mostly from taxes. Hence large tax-collecting agencies exist that have exceptional powers. The spread of electronic cash will eventually give everyone the ability to carry out most of their financial transactions behind a cloak of anonymity that no government agency can pierce.

Any effort governments can make to combat this problem will require money, of which they will be collecting less and less. Once the government gets on the wrong side of this feedback loop, there's no way out of it, except of instituting some kind of totalitarian system.

Few tears would be shed for the government in Silicon Valley, where libertarianism is popular. But in the long run anarchy on the local or global scale is a more serious threat than Big Brother. And it's hard to see what the libertarian approach has to offer on this front. One of the best ways to discourage terrorism is by threatening massive retaliation against sponsor states, but this doesn't work without a powerful central government running a big, sophisticated military.

Beyond that the only way to prevent terrorist acts is through surveillance of everyone and everything. This might seem incompatible with the general anti-government trend. But terrorism is deeply disturbing; anything that appears to combat it, is reassuring; and the citizens of a libertarian Utopia may one day eagerly accept such surveillance. Having failed to sneak in through the back door, Big Brother may return via the front and be welcomed.

Crypto may offer at least one solution to this ultimate nightmare. There is a cryptographic protocol called secret sharing, which is a way of dividing a piece of information into several pieces, which are called shadows. The shadows can be distributed among several recipients. A single shadow or several combined are unreadable gibberish. In order to reconstruct the original information, the shadows of all the recipients must be put together.

Now imagine a digital-video camera on every block, directing its output into a tamper-proof chip that divides it into several shadows. One shadow goes to, say, the local police department, one to the local block-watch group, another to the Civil Liberties Union. As long as any one of these groups withholds its shadow, the information cannot be reconstructed; it effectively does not exist.
It's as if there were no camera. But if a crime takes place in view of the camera--let's say the terrorist puts down his nerve-gas sample in front of it--then the groups can combine their shadows to re-create the original video and catch the terrorist in the act. Implemented on a wide scale, this could make it very difficult for a criminal to get away with a crime. And as long as at least one shadow holder is responsible, the privacy of citizens would not be threatened.

As new technologies continue to appear, we will probably see a lot of initial enthusiasm, followed by a rush of anxiety as we realise how evil people could use them, and then a gradual relaxation as we come to understand how the technologies could also be used to thwart the bad guys' schemes.

Science is an eternal process of consuming and destroying its own dogmas. This can work only where the free and open expression of ideas is fostered. It is no coincidence that science has flourished in the freest countries and that totalitarian societies eventually lost their scientific edge as new ideas were put through an ideological filter.