It's been an exciting week in Second Life! Dell opened a new store in-world, IBM's CEO, Sam Palmisano, made an in-world appearance, and so did the dreaded CopyBot. The presence of companies like IBM, Dell, Reuters and many others in Second Life shows that there is growing interest in using virtual worlds for more than killing orcs and avatar-based flirting. Sure, these companies are just beginning to experiment with business applications of virtual worlds, or "v-business," but things are moving fast. It seems like every week there is an announcement that another new company or institution is trying to go virtual by buying an island in Second Life.
But also this week, the now infamous CopyBot reared its ugly head in Second Life. Here's a brief recap: CopyBot is a tool that enables the unauthorized copying of virtual objects by a player (see Raph for more detail). Since virtual objects in Second Life are created by other players, rather than by Linden Lab, there was an outcry from many players to stop the use of CopyBot. Players protested and closed their in-world stores in fear that their creations would be stolen, resulting in the loss of real U.S. dollars.
In response, Linden Lab banned the use of CopyBot under its Terms of Service agreement but at the same time stated a reluctance to enter an "arms race" with players (e.g., by implementing Digital Rights Management technologies like Apple did with iTunes). The Lindens are taking the position that unauthorized copying is a general problem with the internet, "Like the World Wide Web, it will never be possible to prevent data that is drawn on your screen from being copied." Thus, player/creators in Second Life are panicking just like the Recording Industry Association of America did a few years ago when it began suing individuals for the unauthorized copying of music.
So what impact, if any, might the unauthorized copying of digital content in virtual worlds have on the still nascent v-business industry? Will CopyBot and its successors scare off some early would-be v-businesses?
On the one hand, if companies like Dell use their virtual store fronts primarily to sell physical products, like XPS laptops, unauthorized copying may not pose any problem at all. On the other hand, to the extent that companies attempt to sell digital products - virtual houses, avatar hair, code, images, books, music, etc. - their revenues will be vulnerable to digital piracy. Should companies be worried?
Bob >Linden Lab banned the use of CopyBot
secondlife.com states (on it’s front page): Second Life is a 3D online digital world imagined, created and owned by its residents.
What business has Linden ‘banning’ anything, it does not say anywhere in the above ‘and governed by Linden Lab’, if this is the case, could they put it on the homepage please just so we can be clear what the 'game' is.
Bob >So what impact, if any, might the unauthorized copying of digital content in virtual worlds have on the still nascent v-business industry? Will CopyBot and its successors scare off some early would-be v-businesses?
It might scare off a few brands but until anyone does ‘serious’ business in-world or uses it as a transactional portal then it probably will have little effect, however it will slow the evolution to that point though this goes hand in hand with requiring a critical mass of residents (not ‘players’ btw).
What I mean by these latter two cases are instances where a brand is investing any value into the world that they think, if harmed, will impact their brand in the off-line sphere and / or is something that they have invested enough money in to need to get a return on investment rather than just doing quick ‘n dirty virtual PR, which is what I see right now in most cases.
By transactional I mean using actual in-world code as a key part of financial transactions e.g. to buy a new PC, if the objects that form part of the transaction system can simply be copied or hacked, then there are issues, but I’m not aware, again, that anyone is using SL for that sort of thing at anything beyond a hobby scale.
Posted by: ren reynolds | Nov 18, 2006 at 06:42
There are two distinct steps two copying a virtual object. One must first retreive the characteristics of the object from the system, and then replay this data back into the system's object creator. Preventing the first step is, of course, impossible, since the client needs this information in order to render the object. Obfuscating the protocol (or, as some angry denizens have suggested, bring criminal charges against libSL's developers) is not a long-term solution and, as Linden Lab's recognises, merely the beginning of an arms race.
What I haven't seen addressed in dicussion of CopyBot, is whether technical measures can be implemented at the second step (creating the new copy from captured data) to prevent illicit copying. This hinges on the question, do Second Life object have value outside of Second Life? Obviously I could take a texured model from Second Life and use in another game (perhaps even another virtual world), but I propose that principally player's want objects to use in-world and copying in-world is what most bothers "v-business" owners.
A virtual world could conceivably prevent the creation of new objects identical to an existing copy-protected objects (a hash table, if you will). Simple hashing could be defeating by introducing minor variations into the copies, but it shouldn't be infeasable to develop a domain-specific hash which reduces minor variations on an object to a single hash value (for example, reduce the vertex coordinate resolution prior to hashing to defeat minor peturbations in the vertex coordinates). The system would only need to be strong enough to make any copy visibly imperfect to substantially protect the interests of object sellers and preserve the demand for "official" copies.
At least, I think "data to allow residents to compare asset creators and creation time" is rather weak in comparison.
Posted by: Robin KAY | Nov 18, 2006 at 07:37
The thing that has stunned me about this, with all due respect to Linden Labs, is that such a forward-thinking, metaverse-aspiring, MUD-schooled company could have a "surprise" Napster moment. Granted, the execs at Linden Labs seem to have been well aware of the possibility of a copybot due to the structure of the client-server system, but I would have thought that their official reaction to software that allowed interface circumvention and object duping would have been something discussed and decided pre-launch and spelled out in the TOS and explained to players.
But, okay, LL seemed to have made a vision misstep and fessed up to it very well, so moving past that...
The player outcry at the moment is readily understandable by seeing how players have investments and personal business models built on expectations that, in Second Life, duplicating virtual objects outside of the player permission structures would be not be easy. And I can't really blame the players for not anticipating a popular duping tool -- in most gamelike MMOGs, after all, a lot of effort goes into making sure that player duping is made difficult, for obvious reasons. So players invest under an expectation that duping will not happen to them, duping turns out to be a serious threat (in their view), they see a risk of loss, and they complain about this to the proprietor -- all perfectly reasonable.
So moving past the player reactions, what does this mean for a virtual economy.
Well, I'd like to suggest that it depends on what kind of virtual economy you want. If the virtual economy you want is the typical MMOG economy of the type Ted and Julian and Josh and many others have expounded upon and celebrated, then you want artificial scarcity in the system because, as Ted says, the strange rule of all this is: scarcity is fun. So, better for SL to follow the path of other MMOGs, to please their shop-keepers, and go the DRM anti-circumvention route, building scarcity into their code more effectively.
Somehow I don't think LL is going to be enthusiastic about that, given what I know of their vision -- I could be quite wrong about that, given that outlawing of Copybot.
Now, there's another possibility for a virtual economy -- a world without scarcity. Ted and Richard (and others I'm sure) have warned us that such a world is boring. This is the corollary to "scarcity is fun" -- "no scarcity is not fun." However, I'm not sure I agree that this would be true in Second Life. Dan and I have written about amateur-to-amateur creativity, I've written about non-proprietary information sharing and reputation economies, and Dan has written about the dangers of Walled Gardens in scholarship. I believe you could build a rich virtual and social environment that would generally celebrate, rather than condemn, free information flow and duplication. We have something very much like that now with the World Wide Web, don't we? And SL, I thought, sort of aspired to be something like the 3D Web.
The problem here is that a no-scarcity virtual economy creates an entirely different set of values, strategies, and policy questions than a high-scarcity virtual economy. One dynamic, that I mention in the Digital Attribution paper, is that we see a shift to an advertising model. Creation of content is fueled by promotional dynamics. Many people (including lawyers) feel differently about advertising than they do about other forms of information, as Ellen Goodman has recently pointed out. So, e.g., in the digital environment, the paradigmatic low-scarcity economy problem is spam. Proposed solutions to spam are often about reintroducing artificial scarcity by, e.g., placing a technological "tax" on email communication.
I've gone on a bit.
Raph says that this is about bone v. blood and that Second Life is at a crossroads. I guess my concluding thought is that it would be better if Second Life really thought this out and decided what model it wanted to follow and promote -- there are three that I see:
1) The high object scarcity model that the shop keepers have come to expect and are demanding.
2) The low object scarcity model that is in some ways more in line with their vision, but raises a whole new set of issues and problems.
3) Walking the middle line, pointing to DMCA notice and takedown provisions, tweaking the TOS slightly case-by-case. This approach suggests that SL is going to be just like MySpace or YouTube or any other Web 2.0 play. I could see why SL would want to do this, but this "we're not making out new rules" approach is exactly what was lamented by Yochai at the first State of Play.
Interesting times.
Posted by: greglas | Nov 18, 2006 at 08:46
One clarification:
When Yochai Benkler suggested that Second Life should try to improve on the default IP regime, I think the response from Second Life was that it would, by enabling individuals to choose what regime they wanted. It was a nice libertarian deflection -- "we don't want to create an umbrella IP policy -- the users can create it."
Problem is, as shown by Copybot, ceding power to the people and stepping completely out of the equation invites technological anarchy. Even in the paradigmatic libertarian state, the state performs certain functions, e.g. enforcing private rights of property ownership.
So, applied to the current situation, the thought would be that the shopkeepers have made clear what they believe their private property (IP) expectations to be they want the SL central planning authority to back up those expectations with force.
Point being that the issue that Benkler raised cannot be simply deflected by ceding the IP decisions on this to the users, if the virtual state won't step in to enforce those decisions. SL needs a "vision" here.
Posted by: greglas | Nov 18, 2006 at 09:33
Robin Kay> A virtual world could conceivably prevent the creation of new objects identical to an existing copy-protected objects (a hash table, if you will). <
This isn’t needed in a server based world. No two digital objects are in fact identical in such a world. Each object was created at a different time, and possibly by a different account, and the Server knows that. If objects in a virtual world are not unique, its because designers made a conscious decision not to record that information, presumably to save money. In this post from Robin Linden, it looks like they are addressing the issue, by exposing creation time and account. Like greglas, I am surprised they didn’t have the code in place ready to turn on, when it became a problem. It seems pretty obvious to me, speaking as someone who made that a feature of his own virtual world design years back.
Bob Moore> On the other hand, to the extent that companies attempt to sell digital products - virtual houses, avatar hair, code, images, books, music, etc. - their revenues will be vulnerable to digital piracy. <
I’d argue that “brands” are better protected in a server based world than in the physical one. With the appropriate database entries, I can always be sure I am getting a genuine pair of Nike digital sneakers and not some cheap knockoff. Commodity producers, who rely on selling function rather than origin, have more of a problem.
Greglas> 1) The high object scarcity model that the shop keepers have come to expect and are demanding.<
I’m a big fan of the reputation economy when it comes to digital objects. But you can’t eat reputation. However, if what shopkeepers sell is “the Genuine Article”, they are free to set the scarcity at any level they like. Just as artistic print makers do in the physical world. I think what SL should aim for is a world in which functionality is free, but origin, and the status that goes with having the “genuine article”, will cost you.
As I understand it, most functionality in SL is embedded in code anyway. And an open source, reputation based economy seems quite capable of producing that. They do need to spend some server cycles and database space in keeping the history to support that though.
Posted by: Hellinar | Nov 18, 2006 at 11:21
Hellinar> This isn’t needed in a server based world. No two digital objects are in fact identical in such a world. <
I think you missed my point. You're talking about being able to tell two objects (copy and original) apart based on meta-data. I'm talking about being able to tell when two objects are the same (when an object is a copy) based on physical attributes.
To reiterate: If a client creates a new object which is a copy of an existing one, the system can tell it's a copy because it has already got an identical or very similar object in its database. Hence, you can't make exact copies and any imperfect copies you do make must be changed by at least a certain amount.
My proposition was that unique meta-data (creator, creation date, etc) isn't enough to protect the value of original objects, but that unique appearance might be and futhermore that it is technically enforceable.
Posted by: Robin KAY | Nov 18, 2006 at 12:35
Forgive me for not really knowing all about Second Life outside of what I have read (and there is a LOT to read) but I think that if this virtual world is trying so hard to relate to the real world then shouldn't policing this come from inside Second Life?
In the real world, copyright laws and infrigements are handled by the courts. Maybe Second Life doesn't want to have their world be that "real," but artificially instituting laws through code seems to go against what they are trying to accomplish.
Every game has hackers, farmers and bots. Every one of them. The reason that it is such a big deal in Second Life is because virtual money translates to real money. If there are companies willing to pay people to farm gold in WoW, how far off can they be to figuring out the best way to make money in Second Life? Duping items is just the first step and probably not the last.
Posted by: Iao | Nov 18, 2006 at 17:06
Robin Kay> To reiterate: If a client creates a new object which is a copy of an existing one, the system can tell it's a copy because it has already got an identical or very similar object in its database. Hence, you can't make exact copies and any imperfect copies you do make must be changed by at least a certain amount. <
I think there at least three kinds of copying at work, with rather different implications and solutions.
1) Making a perfect copy, and passing it off as someone else’s work.
2) Making a perfect copy, and passing it off as your own original work.
3) Making derivative works without attribution and/or payment.
Case number one is addressed by server secure history data. Case number two, as you suggest, can be detected by simple hashing. But as soon as you step away from a perfect copy, and in to “very similar”, you are stepping into a area where computers haven’t traditionally been very strong. At least in detecting what humans would call “very similar”.
I’d think attempts to create a unique signature for “similar” objects would be vulnerable to hacking tools that change the generated signature without a noticeable change in user experience. Anyway, such tools are being developed and deployed, so we will have to see how it goes.
Making ‘very similar’ copies shades into make clearly “derivative” copies, and on into “inspired by” copies. Case 3. Which is something I think worlds like SL actually want to encourage. Again, server history is the key here, if you want proper attribution/payment to occur.
One bit of history new creations should carry is a “origins” section, where people can point to the works this effort derived from. And crucially, allow other people to tag in “prior art pointers” where they think a work is an intentional or accidental derivation of things that have been done before. The server can reliably decide if the prior work is actually prior. I’d envisage some sort community process to determine if a work is derivative when that is disputed.
On the whole, I think the outlook for proper attribution of creative works is much better in server based worlds than in the physical one. So long as people treat digital objects as something new in the world, and don’t try to shoehorn them into physical world processes. Unlike the physical world, server based worlds have a reliable historian and watchdog. More virtual worlds should put that to use.
Posted by: Hellinar | Nov 18, 2006 at 18:02
I have no love for Second Life, nor copyright. It is an interesting opportunity to watch these 2 wrestle. I know it is callous to say, but I wonder how many people affected by this are responsible for violating the copyrights of others? How many models of the wares being duplicated were themselves made in a “warez” copy of 3d Studio Max?
These aren't the only factors that will be wrestling in this new exchange, as greglas points out. Will Linden's idealistic Libertarian Utopia be forced to adopt fascist behavior to curb the practice? I am referring to fascism in the classic sense; corporatism. Libertarianism has historically fallen readily on it’s sword at this point in the real world. I assure you though it stops being libertarian when you place the liberties of corporations over those of individuals, and any DRM would be exactly that.
On the plus side this is the most interesting thing that has come out of SL. Primarily they seem only to be on my radar when furries are cyber-sexing and people are paying absurd amounts of US Dollars for virtual real-estate. I would appreciate it if the contributors here at TN would invest time in detailed analyses of all attempts at resolution in these matters.
Posted by: Detritus | Nov 19, 2006 at 09:22
In copyright issues it is often useful to take a very naive viewpoint: If one can create objects without any physical cost, that is, create items of value from nothing, why should this be prevented? Isn't this a sign of living in a paradise?
Of course, the whole reasoning behind copyright is to encourage creation of artistic or at least creative works. But if this is a game, an activity for leisure, it seems to me that people would create these items anyway, for their own pleasure. Is any more encouragement needed?
Posted by: SH | Nov 19, 2006 at 14:41
Personally, I wouldn't mind wearing 'knock-off' virtual clothing that looked the same as the original, or reading a copied digital book that contained the same content as the original. I may even knowingly seek these products out as a way to save some L$ (and therefore real cash)
However, if the proven creator had the ability to delete or significantly modify all of the illegitimate copies, I would think twice about buying such digital merchandise. Imagine buying a book whose content disappeared, or a cyber-sex toy that didn't work as intended. That may convince people to buy the real thing.
Since the use of CopyBot and other similar programs has such significant consequences for Second Life, I'd imagine that Linden Labs will be putting a significant amount of development into ways to automatically detect illegitimate copies.
The real difficulty, as has been pointed out, will be to identify those items which have been slightly altered to try and defeat the automatic copy protection measures. I imagine this will be handled by some kind of community based identification program and the item's metadata (specifically, the item creation time) to identify the original item creator.
The people who create illegal copies, can then have accounts banned in a method similar to how many Games ban Gold farmers, or Virtual Currency brokers. Presumably, this would involve identifying accounts that log in from the same IP Address, using the same billing information, etc.
Posted by: Brandon Checketts | Nov 20, 2006 at 09:34
Hellinar wrote:
This is interesting, and I think metadata & object history can go a long way to protect those who are interested in proper credit for their creations. I also agree with you that the tricky question is the derivative work -- a "unique" object that is substantially based on another's work is a copy in the view of the creator, but from a system perspective, it is probably a new object.
I doubt, though, that your 3 problems of copying jibe with the motivations of many second life sellers, who I think are mostly bothered by the copying per se and not by the potential for passing off. It's not about plagiarism, in other words, it's a complaint about the mere unauthorized "existence" of the copy in SL.
Posted by: greglas | Nov 20, 2006 at 10:23
How are you going to determine whether my model of a VW is based off somebody elses model of a VW..? It is impossible.
The only real solution is to implement two different formats for upload and download (high-level for creation, low-level for rendering) and mutate the download format often. If you allow upload in a low-level format then DRM is impossible (crackers can just capture graphics at the low level graphics driver).
Posted by: Ola Fosheim Grøstad | Nov 20, 2006 at 10:45
>How are you going to determine whether my model of a VW is based off somebody elses model of a VW..? It is impossible.<
While not impossible, its only possible in particular cases. Which is why I was careful not to advocate that. What is eminently protectable in a server based world is originality. Which is where pointers to “prior art” come in. As in the patent field, these don’t assert that you copied my idea, but simply that I had been there and done that before you.
As a designer, proof of originality is important to me. Around 1980, I designed, programmed and marketed the world’s first personal computer spelling checker. In that case, I can reliably demonstrate “first to market”, because there were so few PC magazines at the time. Within months, another personal computer spelling checker was on the market. I am confident this product was thought of quite independently of mine.
Six months later, there was a crude knock off on the market, sold by the owner of a chain of computer stores on the East Coast. I am confident it was a knock off, as he was previously buying my product and it was selling well. Pursuing any action against that however is pretty much impossible for a small business owner in the real world.
In your VW example, if my sales records show you bought a copy of my car on the 2nd of Nov, and on the 5th of Nov you put a similar car on sale, its reasonable to assume you copied it. Mind you, such a blatant audit trail is easy to hide in a virtual world with alternate accounts etc. So I would go with protecting originality. Which I still assert is much easier to protect in a server based world than the physical one.
One thing I found interesting in this related article in Wired is the reference to the high fashion and haute cuisine industries. They reference a couple of papers which demonstrate that though copying is rampant in both industries, claims of originality are jealously guarded. And it’s the protection of originality, rather than the prevention of copying, that promotes the high rate of innovation in those industries.
Posted by: Hellinar | Nov 20, 2006 at 13:10
Yes, Hellinar. Originality is a good measure that can work for constructive virtual worlds. Copying was also rampant among classical composers. Copying can either be viewed as a salute to your design being worthy of building on, or it can be viewed as theft.
The Metaverse-style world suffers from two issues in this regard:
1. The capitalistic spin detracts from the creative commons ideal (theft versus salute)
2. The Real World simulation aspect makes originality harder to determine. Especially if people try to recreate familiar objects.
Agree?
Posted by: Ola Fosheim Grøstad | Nov 20, 2006 at 15:36
@Ola: Good points. An awkward underlying reality though is that digital worlds have little room for capitalism in the classic sense. Capitalists in the historic sense provided the vast amounts of capital to build the factories needed for industrial production, and the distribution systems to deliver it. In a digital world, once the creative designers has done their work, reproduction costs are close to zero. As is distribution cost. No capital required.
Consumer based worlds like WoW do need large amounts of real world capital. The content is produced externally by large amounts of artists chained to their desks by money. But once those same creatives move in world, and take advantage of the low reproduction costs and distribution costs, and “word of mouth” marketing, the requirements for capital are a lot less. Real world dollars can buy you some reputation I think, but not as much as talent, or plain luck.
As to the copying of real world familiar objects, isn’t that usually done with a conscious tip of the hat to the original real world object? What would be the point of your digital VW model if no-one recognized it as a riff on the famous bug? Your interpretation might be original, but not the classic shape. It would be up to Volkswagen to decide if the sue or smile.
Posted by: Hellinar | Nov 20, 2006 at 17:37
Hellinar: It would be up to Volkswagen to decide if the sue or smile.
Yes, but how do you resolve disputes between players? If Ann has spent two weeks recreating a VW in your world, and then see Pete having a similar VW, she would cry "THEFT!". However you can't resolve the dispute based on originality and not based on technical evidence (Pete might have piped Ann's model through a program that recreates the model with a slightly different set of primitives or a different configuration etc, then added some mods to it.)
Anyway, I am not sure if letting users having ownership in the world is a good idea.
Seems like your originality measure could go well along with proclamations and a sense of belonging instead. (E.g. fashion shows, car shows, etc) Maybe that is what you meant with reputation.
Posted by: Ola Fosheim Grøstad | Nov 20, 2006 at 19:19
Sorry for the thread necro here, but this is a really interesting discussion (especially Raph v Prokofy on Raph's blog). Most of it's just a rehash of things I've heard for the last 20 years, but there's a cool twist in Prokofy's implication of Linden Labs.
Is it in the best interest of SL to be slightly dangerous? Does the "real" metaverse need a subculture of predatory techno-elite? Maybe just the possibility of such a subculture to stir our imagination and make us hold our breath when visiting dark alleys?
Posted by: Ken Fox | Nov 29, 2006 at 20:17