There have been several recent reports of high-profile incidents of griefing and even an effective "denial of service" (DOS) attack within Second Life. The worst of these shut down the system entirely due to the behavior of in-game scripted objects. In this incident, a self-replicating object did just that: replicated by doubling itself to the point it brought the servers down. DOS from the inside, apparently intentional, using the tools provided by Linden Labs to its users.
According to reports Linden Labs first changed the capabilities of the scripting language to prevent this sort of thing from happening again, but then rolled back that change due to the outcry from honest SL citizens, as it effectively disabled many of the scripted objects they had come to depend on.
So the ultimate solution? Not content with banning the individuals involved (generally acknowledged as ineffective in SL and other virtual worlds/games, given how easy it is to come back again), Linden Labs has called in the FBI.
The only thing that surprises me about this is that more of this level of griefing hasn't been more evident already Early on in Second Life, Randy Farmer made an invisible repeating auto-cannon that fired a hunred rounds per minute and teleported people away, just to illustrate what could be done with the scripting tools available. He could have been much nastier with his creation, but chose to limit the gun's abilities. Since then there have been reports of other local griefing tactics using scripted objects, but nothing major. Perhaps this latest event is an example of a network effect in action: as SL's user numbers climb, the probability of such "attacks" approaches certainty (ref: Hanke's Law). And certainly the attacks won't be this simple or overt in the future. Will the next griefer-object be a cute Trojan puppy that's given away to people, gently self-replicating only when there's no other copy around (so that you can give one to a friend!)... but which, at a pre-appointed time "explodes" -- deleting your clothes, teleporting your character, changing into a bloody pile of gore, or worse?
So here's the bind: in Second Life, the attraction for many is that you can create whatever sort of functional objects you want. And you own what you create... except you really don't. You retain the intellectual property rights, but not surprisingly Linden Labs explicitly retains the ultimate right to delete anything you make in their world. They also retain the right to change the underlying script language, in effect retroactively changing the physics of the world and what can be done in it.
I'm not by any means taking LL to task for providing such open-ended scripting capabilities, nor for protecting themselves by retaining the ultimate ability to change or delete objects. If nothing else this is a courageous experiment, as it gives users abilities unparalleled in any other major virtual world. But what does this say about the relationship between user-created content, user-ownership of content, and corporately owned virtual worlds? Can these truly co-exist? Can a virtual "country" (as SL likes to consider itself) exist without an operative constitution and in-world laws and enforcement?
Is there a happy medium that lets people be creative and gives the virtual world the enormous benefits of user-created content, but which protects the bulk of the user populace and doesn't leave them open to self-replicating data bombs, pushed porn objects, teleporting auto-cannons, or even more nefarious objects? Or, despite the siren-song of distributed, freely created content, does the risk of user-created content with scripted functionality ultimately outweigh the potential benefit to the user base and the corporate owners of the world?