Are we doomed yet?

New rules for a society of wizards are being proposed and implemented every day. In recent times, a leading technologist has called for us to reconsider "the open, unrestrained pursuit of knowledge," a federal judge has ordered an injunction against hyperlinking with "a desire" to "disseminate," and the NYPD has upgraded its surveillance network. These changes reflect the increased power of individuals -- and the need of governments to keep individual power in check. Ironically, though increases in individual power during the print revolution catalyzed the ideals of freedom and individuality, corresponding increases in individual power during the computer revolution have catalyzed a sense of doom and a desire for autocratic rule.

As the "functional" language made possible by computers increases in scope and power, we will want to police its misuse. The dangers are too great to ignore. However, we do have a choice analogous to the one David Brin gives for surveillance technology. We can attempt to implement a centralized, top-down apparatus to monitor and manage what is said (or "coded") -- or we can use a more open system. The government can try to bottle up dangerous information, or it can yield to the disseminative power of the Web and instead help companies and individuals defend themselves against harmful technology. For practical reasons, the government may have to yield to the Web, but is there any logic in abandoning a preemptive defense? It is a small matter to grant DVD pirates a brief holiday while the movie industry works out a new standard; it is an entirely different matter to let a reckless undergraduate post the molecular structure of a flesh-eating nanobot.

I think the reality is that preventing dissemination on the Web is like exterminating cockroaches. We could put centuries of effort into the problem, but hackers will still scurry out from under the cookie jar. We will find out if that theory is correct during this decade, during which the governments and corporations of the world (at least in the West) will try to stop online piracy. Clearly, if a society can't stop teenagers from spreading the word on how to "rip" DVDs or trade MP3s, then that same society will be hard-pressed to stop terrorists from spreading Ebola-AIDS or the latest self-replicating nanobot.

We might just have to accept that dangerous technologies are coming. Joy recommends "relinquishment" of certain lines of research, but even if the governments and corporations of the world could agree to swallow such a bitter economic pill, the kooks and hackers would continue chipping away at the unknown and publishing their findings on the Internet. Understanding will move forward, one way or another. Getting through this century may be a crapshoot for the human species (Joy cites Ray Kurzweil's "optimistic" prediction that we only have a "better than even chance") but sometimes you just have to roll the dice and hope for the best.

The question is: What sort of society is most likely to make it? If a police bureaucracy could really muzzle 15-year-old Norwegian hackers like Jon Johansen, then maybe it would outlast the alternative. But what happens if one clever kook slips through the cracks? What happens, in a police bureaucracy, if someone releases a nanotech plague into the environment? If the police can suppress information on the structure of the nanobots, then only a handful of government bureaus and hand-picked researchers may be allowed to work on a cure. Millions could die waiting for the bureaucracy to solve the problem. On the other hand, if the molecular structure of the pest is published worldwide, anyone with the expertise could help design defensive technology.

This kind of model has worked well in the fight against computer viruses and worms. The worm that Robert Morris unleashed in 1988, which shut down close to 10 percent of all networked machines, was decompiled within 12 hours, thanks to a spontaneous and collaborative effort between scientists at MIT, Berkeley, Purdue, and other universities. Berkeley released a full patch for Berkeley UNIX 15 hours after Morris released the worm. As an undergraduate that year, I casually came across a complete explication of how the worm worked, including the source code, posted to a public FTP site on machines at the MIT AI lab. The result of this open distribution of dangerous knowledge wasn't a meltdown of the Internet but improved security around the world. The Net got hit by an epidemic, some machines went down, but the system rebounded within days.

In his ruling against 2600, Kaplan compared DeCSS to a "propagated outbreak epidemic," an epidemic in which the disease is contagious, in contrast to a poisoned well, a "common source epidemic" which can be halted at the source. The analogy is apt. The self-replicating threats of the 21st century will propagate from location to location the way computer viruses now hop from computer to computer. The difference will be that instead of wiped hard drives we might have sick and dying people. A couple years ago, for instance, I would have lost my father to the "ILOVEYOU" virus. Perhaps we all can think of friends and family members who would have departed this world if computer viruses had real-world nanotech components.

A grim future indeed, but I am cautiously optimistic for a couple of reasons. First of all, most people in the world, despite their differences, want stable, healthy lives. As we have seen with the Internet, .1 percent of the population may always try to throw a wrench into the machine, but the rest of us will scramble to fix the problem, punish the pranksters, and defend against wrench-throwing in the future. Second, I think that even among the pranksters only a very few will cross over from fun-and-games with computers to deadly real-world viruses. At the worst, we face a few crazies and, more seriously, a handful of "rogue" nations and terrorist groups.

But it only takes one crazy, in theory, to invent a new disease. We might ask whether any society at all, free or totalitarian, could reverse-engineer a viral or nanotech pathogen fast enough to create a "cure" before the population is decimated. The example of nuclear weapons shows us clearly that a technology can be easy to deploy offensively but nearly impossible to defend against. This may be the situation with self-replicating threats, in which case we are doomed. Even so, our best chance of survival lies not in criminalizing certain kinds of expertise or knowledge but in disseminating that knowledge as widely as possible, so that any attack will be met by the widest possible resistance, a citizens' militia of knowledge workers, rather than a handful of cronies in an intelligence agency.

Classified knowledge creates divisions and hierarchies. In Deus Ex, the classifications in a U.N. intelligence agency run parallel to the levels within a secret society and take their nomenclature from the angelic ranks. The player moves through a world patterned by the natural law doctrine of Thomas Aquinas, in which the average citizen must accept the divine right of informed rulers, who through their access to information remain closest to the mind of Bureaucracy and therefore to the source of legitimate power.

Though we might be foolish to put too much faith in the romantic notion of the "citizens' militia," we should be very suspicious of laws that limit the creation or dissemination of knowledge. They threaten to create a privileged class of information shepherds who, though well-meaning at first, could easily abuse their dramatic power advantage over information consumers. We should not give up our freedom to know and to communicate unless we are certain that the new order would be vastly more secure than the present one -- and, as I argue above, the likelihood is that it would not.

- - - - - - - - - - - -

Knowledge hierarchies exist in every society and will continue to exist. The better educated will always rule the less educated. Modern democracies rely on public education to make this division somewhat permeable, and such systems are probably our best hope for preserving "liberty" for the average citizen. As the essential qualities of knowledge change, however, so will education. The fundamental threat lurking behind any reevaluation of knowledge is much broader than a potentially corrupt intelligence agency. The threat is that public education might become increasingly circumscribed, strengthening class barriers rather than softening them.

Instead of regulating access to the engineering languages of the coming century, we should teach the languages to anyone who wants to learn. We already see a dramatic economic divide between the computer literate and illiterate, a divide that will only widen as more products and technologies are developed on-screen. To say that learning to use a computer is like learning to use a new language is a metaphor but also more than a metaphor. Building anything with a computer -- from software to circuits to research papers -- requires the composition of logical and symbolic elements, abstracted through software. Some applications more closely resemble human language than others, but to be mastered they all require the same symbol-processing ability.

This is why we should be concerned about revisionist educators like William Crossman, who believes that we should forego teaching the masses written language and instead give them access to the world's knowledge through the magic of VIVO (voice-in, voice-out) computers. In his book, "Compspeak 2050: How Talking Computers Will Recreate Oral Culture by Mid-21st Century," he predicts that all language functions other than speaking and listening will be handled by VIVO computers by 2050.

"Like most technologies, written language will serve its function until some better technology comes along to replace it," he writes. He believes that oral-aural communication is more efficient than writing and reading, and this is the reason, he says, that the average person prefers the telephone over the letter and the TV over the book. "In the 21st Century, people with access to VIVO-computer technology will once again be able to use spoken language to access all stored information. Talking computers are going to make writing, reading, spelling, alphabets, punctuation, written numerals, music notation, and all other notational systems obsolete."

As a long-time user of voice-recognition, my patience beaten down by the number of times every program I have ever used has confused "of" with "up," or "was" with "wasn't," or "their" with "there," I am certain that his expectations for the technology lean toward the optimistic. Nevertheless, even if recognition algorithms are perfected, we won't ever abandon notational systems. They are essential for organizing any large project. Today's electronic entertainment certainly contains many "oral-aural" elements, but without exception it is created with writing, editing, storyboarding, programming, and other notational systems.

This will continue to be the case, even if the mass of consumers becomes illiterate. In Deus Ex, conceivably, if we had tweaked one or two minor features, we could have released a game that required no reading ability whatsoever; a player could have walked around the 3-D world, listened to newspapers being read aloud, watched his character enter door-codes automatically, and so on. The average consumer might well have welcomed a game interface that required no reading. But it would have been impossible to make the game without one or more scripting languages. Even the dialogue, written in a proprietary "conversation editor," was inseparable from flag-checks, flag-setting, definition of camera angles, and other game logic, all integrated with the speech of the characters. Without explicit notation for game-state, this logic would have been opaque to the developers. A VIVO computer might have been able to convert the various "if-then" verbal constructions of the six designers, three programmers, and three writers into some sort of internal format, but without a common table of flags, defined at least with pictographs, the workflow would have been chaotic and debugging would have been futile. The situation would not be so different if we were stacking shots of a movie script to hand off to 100 CGI artists.

In other words, the average consumer of entertainment can easily be excused from learning to read, but those in the business of producing it will always need a chirographic means of managing assets, workflow, and logic. In a similar vein, the average consumer of products can be excused from learning to read. Within a few decades, VIVO software could conceivably handle a consumer's checkbook, taxes, VCR programming, warranty claims; etc. The average high-school educated person is only required to read in a handful of situations as it is. Those annoying New Patient forms we have to fill out at doctor's offices? They could be eliminated, for all of us, with card-swipe technology that has existed since the '80s. Crossman is right: no one will need to read in a few decades, if we engineer the right software. However, the developers of products will require the same written skills as the developers of entertainment, and businessmen are not likely to abandon written contracts any time soon, even if they begin accepting voiceprint signatures. Home-buyers of today often let their eyes skip over the myriad fine-print details of mortgage contracts, and willingly yield to the verbal explanations of the realtor sitting at their elbow, but bank officials will always pore over every phrase with great care.

So, excusing the next generation from learning to read and write will succeed only in sharpening the divide between rich and poor, producer and consumer. If we return to the question of how to safeguard against self-replicating threats, however, maybe Crossman is on to something. Mass illiteracy would reduce the bulk of humanity to a herd of wait-staff and bus drivers, who would be easy to police by conventional means. Resources for truly transparent surveillance could be concentrated on the minority who receive a "dangerous" education. The Jon Johansens of tomorrow would be much easier to spot and guard against. We would have a clearly defined elite, not unlike the Party in "1984," and they would have to meet rigorous ideological and behavioral standards in order to keep their privileges.

Sound like a fanciful projection? This is David Brin's "City Number One" re-imagined with a broader set of technologies. We do in fact face a choice of two cities, and the one I just described is not so different than any future we might imagine in which our fundamental attitude toward knowledge has changed. Without a doubt, it would be the safest solution to problems being considered by Bill Joy, the U.S. government, worried citizens like Schweitzer and Dorsch, Judge Kaplan, and others. Limit education and you limit dissemination. If language will soon acquire the ability to script and compose matter, then limit access to language technologies and you nip high-tech terrorists in the bud.

Universal education has been a brief experiment in the English-speaking world, and there is no reason to think it will stick around forever. Free education for all children became available in England only in 1899. U.S. children have had that luxury only since the 20th century, and we shouldn't forget that even the Enlightenment thinkers who laid the foundations of democratic ideology had a sweeping contempt for the "canaille," or "common people," and were very pessimistic about the prospects of educating them. If we allow our basic attitude toward knowledge to shift -- if we get in the business of criminalizing, censoring, monitoring, and limiting various kinds of knowledge -- I believe we will very quickly slip away from the ideals of universal education, open scientific enquiry, entrepreneurism, equality of opportunity, and the fecundity of creative effort that has made Western democracies so strong during the past two centuries.

The self-replicating, scriptable technologies are here and still arriving. Progress will continue. We aren't choosing whether or not to eat from the Tree of Knowledge. We are deciding whether to put a fence around it and ration the fruit. The choice is not between a perilous freedom and a secure tyranny, but rather between fear and trust. We might even find it easier to trust one another if each of us takes a bite out of that genetically engineered FLAVR SAVR tomato and gains the same knowledge of good and evil.

• View blog reactions