Legal Affairs
space

space


To Kill an Avatar

Norrath, the online world created by Sony, has more residents than Miami and a bigger GNP than Bulgaria. Who will make its laws?

By Dan Hunter and F. Gregory Lastowka

ON YOUR COMPUTER SCREEN YOU SEE YOURSELF APPEAR as a toned and taut female figure dressed in a shiny black half-shirt and a pair of tight-fitting Levi's with distinctive stitching. Your stylishly coiffed head turns from time to time, your big blue eyes blink, and your shoulders rise and fall with your every breath. Behind you stretches a sandy beach and a crystal-blue ocean. You hear the pounding surf and see a cluster of cartoonish grass huts sprouting in the distance. Alice and Leet are standing nearby. You walk over to them and notice a speech bubble hanging over Leet's head as he tells Alice a joke. She cocks her head with interest and then laughs out loud. Both of them see you approach and they turn to you, smiling. Leet waves hello and asks whether you'd prefer to hoverboard or race dune buggies. "Dune buggy," you type. Alice yells "Yay!" and jumps in the air.

Though you know you are here, sitting at your computer, you are, virtually, There. Unveiled in January, There is an online place where you can hang out, chat with friends and family, and go dune-buggying. It is one of many virtual worlds where millions of people spend large portions of their lives. (A study conducted on another world suggests that the median user spends about 20 hours per week in the game.) Some worlds, like Ultima Online or EverQuest, are Tolkienesque games where you might be an elf or a dwarf fighting against dragons and goblins. Virtual worlds such as There or The Sims Online are modeled on more familiar environments and seem like three-dimensional electronic chat rooms where each friend on your buddy list happens to look like a digital version of Jennifer Aniston or Tom Cruise. (Think AOL, but with wish fulfillment.) This digital you is called an "avatar," a term borrowed from the Hindu religion that otherwise connotes the incarnation of a deity. Through your avatar, you move around the representational spaces of virtual worlds, seeking out interesting new locations, acquiring and creating virtual objects, and meeting other avatars. Graphics-based virtual worlds are extremely beautiful, with majestic sunsets, delightful beaches, even dual moons. And like real places in the real world—Philadelphia, say, or Rio de Janeiro—they continue to exist and evolve when you are not around.

While reliable indicators are hard to come by, it seems that the average player of a virtual world is a man in his mid-twenties who earns around $30,000. Access to real-world cash is essential, because most virtual world players pay subscription fees of about $10 a month. EverQuest, introduced by Sony in 1999, now collects $12.95 a month from 430,000 subscribers (on top of the $21.99 price of the start-up software). Though Sony's initial investment in creating EverQuest's virtual world was significant, the subsequent costs to operate it are modest, with expenses mostly limited to maintaining the servers, expanding and refining features, and providing customer support. It's not hard to see why the business of creating and running virtual worlds is popular. Outside of pornography and eBay, virtual worlds are one of the few online businesses that are making money.

KEEPING A WORLD PROFITABLE means keeping its subscribers' monthly payments coming in, which means keeping subscribers happy. Like political leaders in the real world, game designers are under constant pressure from a citizenry with competing demands. Should a game stress equality of power among avatars or should it reward with greater power those who invest more time in the world? Striking the wrong balance could make the virtual world less attractive to new users or offend the subscriber base of long-time residents. And designers must be mindful of the principle of liberty as well as equality. Part of the appeal of the virtual worlds is the feeling of freedom they offer users, allowing them to interact with others and shape their environment. Each restriction on avatar freedom and power may undermine the allure of the virtual environment.

One problem that has troubled virtual worlds from the start is crime. It isn't obvious why online communities have problems with crime at all. You might think that in worlds where avatars can fly, spells can turn pumpkins into coaches, and it's possible to custom-build your own virtual castle in the air, the incentives for criminal activity disappear. But greed and schadenfreude have carried over from the real world to the virtual, whose residents demonstrate a propensity for all the familiar crimes and social frictions that characterize the real world.

Take for example the first case of fraud in There, in which an avatar put up a "For Sale" sign in front of a house that he didn't own. Since few users knew how property transfers were supposed to work in this brave new world, the scam artist collected some serious Therebucks (the currency of There) before the creators of the world discovered what was afoot and took corrective action. This was not just a case of imaginary theft. Therebucks, like most virtual currencies, have real-world value. The economist and virtual-world theorist Edward Castronova made a number of calculations regarding the economy of Norrath, the virtual world of Sony's EverQuest. Based on trading indices from eBay and elsewhere, Castronova was able to calculate the de facto exchange rate between the currency of Norrath and the U.S. dollar. His findings suggest that in 2000, the GNP per capita of Norrath—the total, in real dollars and cents, of all goods and services produced there—was roughly equivalent to that of Bulgaria.

In addition to problems of fraud, virtual worlds are plagued by virtual violence, theft, and harassment. One early incident of violence occurred in the text-based virtual world of LambdaMOO. Julian Dibbell gave a first-hand report in The Village Voice of how two female avatars were "raped" by a male avatar whose owner used a coding trick to control them and then sodomized them in a public room before a large number of others. Though the only loss inflicted was of dignity, the incident had distressing consequences for the victimized avatars' real-life controllers. Dibbell famously dubbed the crime a "rape in cyberspace."

More prevalent than rape is the crime of player-killing. In virtual worlds where avatar violence is possible, more advanced and powerful users sometimes harass, torture, and kill the hapless avatars of neophytes, rather than going after the ogres and orcs that the game developers have created as appropriate sword fodder. The death of an avatar will sometimes terminate the avatar's existence entirely, but this is not usually the case. In most virtual worlds, death is reversible: An avatar can be reincarnated, resurrected by a healer, or otherwise brought back to life. This process, however, is often attended by other serious penalties. For instance, in Lucasfilm's Habitat, one of the first virtual worlds, death was expressed by making the avatar reappear decapitated at his home, with all of his possessions lost at the scene of the crime. The avatar would then need to find a way to reattach his head—and rebuild his virtual life from scratch. What may look like murder on the computer screen may in fact be closer to real-life assault, but it is hardly a victimless crime.

Stopping such crime, or at least containing it, has proved to be one of the most vexing challenges of maintaining an online community. If paying subscribers are constantly being slaughtered and robbed by avatar miscreants, subscriptions will surely decline, hurting the bottom line of the world's owner. In Lucasfilm's Habitat, citizens complained to the company that player-killing and corpse-looting were detrimental to the game's future. The designers' initial solution was to banish death within the city limits by coding it out of the program. Stanford Law School's Lawrence Lessig has memorably advanced the idea that code amounts to law in the technological realm of cyberspace. Nowhere is this analogy more literal than in virtual worlds, where software designers provide the law, the courts, the constitution, and the very physics of existence.

IT IS THE NATURE OF A SOCIETY AND THE NATURE OF ITS DISPUTES that determines the regulatory response of the state. Virtual worlds are little different from real-world societies in this regard, and their regulatory options range from heavy-handed crackdowns to the promulgation of social norms. Unlike the real world, of course, virtual worlds are representational creations constructed of human-written code that designers can manipulate with uncommon precision. For example, in There some users delighted in driving their dune buggies into groups of chatting avatars, scattering them like tenpins. Designers added a "forcefield" option to each user interface. With your forcefield turned on, the disruptive dune-buggy-driving "bowler" just bounces off you.

But such code-based efforts by virtual-world owners are not always successful. Consider what happened to Ultima Online, where designers attempted to address the problem of player-killing by instituting a system for publicizing criminals' pasts that operated much like Megan's Law. The game designers modified the program so that each time an avatar killed another player, his name—normally displayed on other players' screens in blue—would glow in a redder hue, broadcasting his violent tendencies. New players were advised to "read" the reputation of other avatars and understand the dangers of speaking with those who weren't "pure blue." To allow for the possibility of rehabilitation, they coded a way for an avatar to reclaim his "blueness" by attacking known player-killers, creating a loophole that was immediately exploited. Player-killers paired up, attacking each other as a way of restoring their good names. The experiment in fostering better living through technology was abandoned, and the designers chose instead to go the Habitat route of making killing impossible except in certain zones.

Code is immensely powerful, but it is not a panacea for virtual crime. Certainly, if one were to virtually bind and gag all avatars, virtual crime would disappear. But, for a business dependent upon increasing subscribers, this would be suicide. Designers of violent worlds such as EverQuest or Ultima Online don't have the option of banishing virtual death completely—the thrill of risking "death" is part of the game's pleasure. Moreover, the incentives that create virtual player-killers do not seem to be based upon a rational calculus. Coding player-killing to make it less profitable will probably not solve the problem. The narrator of Johnny Cash's song "Folsom Prison Blues" famously shot a man "just to watch him die." A virtual rogue is probably cut from the same cloth. Dibbell's virtual rapist defended himself by saying that the whole thing was just a game and that his gruesome and sadistic virtual acts were simply an interesting experiment.

If avatars find it amusing to make the lives of others miserable, they will find ways to do so. Designers and coders can't anticipate every possible way of circumventing coded restrictions on behavior. Consider the problems with online speech in worlds such as The Sims Online and Dark Age of Camelot. Participants overwhelmingly demanded that their avatars be given the freedom to speak with each other. But once real-time chats were allowed, designers quickly realized that it was nearly impossible to develop an algorithm to block offensive and abusive speech. Technological filtering proved largely futile because offensive phrases such as "ph*ck_y00, 4ssh0l3!" pass cleanly through an algorithmic profanity filter, even though humans won't have much trouble gleaning the message.

Alternative regulatory approaches have been tried, with equally unsatisfying results. To address the problem of harassment, designers have adopted real-world legal solutions such as end-user license agreements. Players are required to agree not to "grief" (harass) other players, and designers are given the option to revoke the subscriptions of those who pester or harass others. But this kind of monitoring is, for the most part, difficult to do: Designers, like federal prosecutors, have overbroad laws but not the resources to enforce them. As in the real world, online policing costs money, and it drives up the cost of subscriptions. Sony doesn't want "law and order" from EverQuest, just profits.

WITH DESIGNERS OFTEN ABDICATING RESPONSIBILITY, virtual inhabitants are left to solve problems on their own, and the mutability of online identities has made some novel solutions possible. Like real-world harassers, online sexual predators seem to prefer targeting women. But gender in the virtual world is an attribute to be chosen, and some women choose to avoid griefing simply by foregoing their biological sex and playing as male avatars. Similarly, male players cross-dress and defend gender swapping as a move that makes strategic sense. They argue that while female avatars are harassed more often, they also obtain numerous benefits, like free gifts and favors from male avatars.

Other social responses are more traditional. In Habitat, players instituted a virtual church to promote the concept of avatar nonviolence. Other virtual communities, such as those in End of the Line, enforce norms against player-killing by putting player-killer names on a board with posted rewards. Posses are deputized to hunt them down, execute them, and confiscate their property.

Grassroots activity of this sort has its limits. It should come as no surprise that the owners of the current crop of virtual worlds—Sony, Microsoft, and Electronic Arts—are not rushing to create political spaces, nor do they wish to identify their virtual dominions as nascent democracies. But democratic alternatives are possible—at least in theory.

LambdaMOO was created in 1990 by a researcher at Xerox PARC as a social experiment. Its designers voluntarily relinquished control over social regulation, implementing an elaborate petition system instead. Anyone could petition for a change in the social structure or architecture of LambdaMOO. So there were petitions seeking the institution of new property rights, a ban on bulk e-mails, and a rule against confusingly similar avatar names. (There were even petitions seeking the removal of the petition system.) Once a petition got enough signatures, it went up for a vote by all users, and the results were then published. Once new rules were approved, the game's wizards (as its designers were called) implemented them using code. Still in existence today, LambdaMOO stands as the most inclusive and democratic virtual world.

Yet LambdaMOO's text-based environment pales in comparison to the visually immersive worlds of There and EverQuest, and few people participate in it. Like all of the most popular virtual worlds, EverQuest and There are the children of large corporations; their massive visual environments require vast resources to build and maintain. At times, it seems like LambdaMOO survives mainly to allow endless theorizing among sociologists and political scientists about the emergence of society in online environments and about how virtual worlds might give rise to unique and innovative forms of politics and law.

Other political pundits contend that virtual worlds make a strong case for limited government. In a recent article in the online magazine Slate, Robert Shapiro, an undersecretary of commerce in the Clinton Administration, argues that a system of sound property rights and initial equality in the virtual world EverQuest—all characters are created with nothing to their names—has allowed limited government to flourish and "economic nature" to take its course. "Real equality," he concludes, "can obviate much of a democratic government's intervention in a modern economy."

SHAPIRO TAKES THIS ANALOGY A BIT TOO FAR. Unlike the designers of EverQuest, Congress cannot create the conditions necessary for limited government by, for example, coding away poverty and sickness or letting us all choose our gender and race. And while the markets of EverQuest may work as smoothly as real-world markets, the politics of EverQuest do not appear, to the casual visitor, as any to which a real-life government would actually aspire. Norrath is brutally violent and anarchic. Its world may be fun, but it is dominated by the invisible yet powerful hand of the Sony designers. Rather than trying to extrapolate ideas about real-world government as Shapiro does, it seems wiser to study the law of virtual worlds as a largely sui generis field.

It is important to understand what role law is playing—and should play—in virtual worlds. Over the next few years, increasing numbers of communities will come to exist in virtual worlds and real-world economies will continue to bleed into their virtual counterparts. (Companies like Nike and Levi Strauss have already agreed to set up shops in There, selling virtual-world products—and promoting real-world ones—to virtual customers). The types of rich social interactions that can now be found in places like There will one day be built into all communication systems. Why would you want to talk with a distant friend on the telephone, when you could meet her in a lush tropical forest online? But keep in mind that Microsoft, Sony, and other companies own these environments. Our lives in the virtual world will not be structured like our current real-space world of sidewalks, parks, coffeehouses, and living rooms. This is because, contrary to what Shapiro says, the government of Norrath is not limited. In virtual worlds, where code is truly law, the designers of the world are that world's de facto authority. Because they are not held accountable to any democratic process, virtual worlds are dictatorships of the most absolute kind.

Yet because their "subjects" are free to subscribe to a competitor's service at any time, these dictatorships do not pose the same problems as the traditional kind. And in any event, "good government" may not be what their subjects desire. To the extent that good government aims to improve the commonweal, good government is exactly what most virtual worlds seek to avoid. In such worlds, scarcity, danger, and struggle are features, not bugs. Programmers could have easily created virtual worlds where every avatar had the option of obtaining every available power. That was tried, very briefly, and people found it boring.

If EverQuest's substantial numbers are any clue, people seem to prefer to suffer through adversity in order to "earn" their status within a virtual community. The resultant struggles among avatars to increase disparities in status, wealth, and power are part of the entertainment. When it comes to the virtual world, it is the will to power that is the feature, and equality the bug.

Because adversity and violence are features of these worlds—and not just side effects—it is wrong to think of crime as an unfortunate byproduct of virtual societies. Crime is simply an extension of many of the same freedoms that make these worlds so appealing to their inhabitants. Game designers have their hands tied. They should, and do, take significant steps to enforce basic rules of fairness, trying to hand out power equally at first and to create impartial systems by which power can be obtained. But if they allow avatars the freedoms to shape virtual society, they cannot clamp down on the many ways that avatars use these opportunities and powers to commit virtual crimes.

You could make a virtual world without the possibility of crime—but it would probably be about as dynamic as Pong or Tetris. It turns out that as we build denser, more immersive, and more compelling virtual realities, we bring into our virtual realities numerous unanticipated real-world potentials. By creating virtual lives, investments, and freedoms, we create the conditions for virtual crime. Is there a solution? Short of changing human nature, there is probably no way to avoid the difficulties of crime, at least if we want our virtual worlds to be as engaging as the real one....

Dan Hunter teaches at the Wharton School, University of Pennsylvania. F. Gregory Lastowka is an associate at Dechert LLP in Philadelphia.

space