Virtual epidemiology
Aug. 21st, 2007 10:55 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
mirrorshardThe BBC have finally picked up on the World of Warcraft 'corrupted blood' incident, previously discussed by the game-studies and game-design community here. (Yes, that's almost a two-year gap between the timestamps on those posts. That's the BBC for you.)
To be fair to them, they're talking about it because health researchers are starting to think about virtual worlds in the context of epidemiology, and whether they'd provide a better model for studying the spread of infectious diseases than standard computer models.
There are a lot of unexamined assumptions there, though, and it sounds as though this team really do need a virtual world designer or theoretician on board. The first big one that jumped out at me is that there was quite a lot of griefing going on, with many players deliberately infecting others. The pat answer to 'why is this a problem for the model?' is that this behaviour doesn't happen to nearly the same extent in real-life epidemics (I'm sure it happens in some cases), but there's a deeper issue as well.
Deliberately infecting others involves sacrificing your character's life - that doesn't mean much in WoW, but the important point is that it means very different things to different classes of player. (If you want to go into this in more detail, the reference is Richard Bartle, and his matrix of player types.) And I'm not at all sure that this level of heterogeneity, of differing levels of commitment to (investment in) the world, translates at all to the real world.
The second big one is that the ethics of the case are problematic to say the least.
So, of course, it would be entirely ethical and uncontroversial to release an infectious disease in a virtual world, without explicit opt-in or any of the usual niceties. Really it would. And the other problem, of course, is getting a virtual world with a decent population to do it in. You certainly won't attract anything near the required numbers (ie. large enough so you can study them statistically with confidence, and do the same with subgroups - large enough so they can be thought of as a fluid rather than a granular, er, blob) just by wanting sign-ups for an academic study. You won't be able to do it in an existing large VW without i) cutting a deal with the designers, who will not unreasonably want to know what's in it for them, and ii) risking some serious legal problems. Different jurisdictions treat virtual property (and virtual people) very differently, and if a third party is putting them at risk then it's very possible that that third party could suffer some consequences. It's a near certainty that they'll suffer from the process, even if found in the clear.
There are basically three levels of analogy on which we can think about the digital Black Death experiment (which hasn't been done, and I suspect won't be). The first is the deliberate spread of a plague affecting humans. The second is the deliberate spread of a plague affecting people's farm animals or pets (foot & mouth disease, for instance), ie. non-sapient (but still sentient) dependents with their own concerns distinct from their human's. The third is a botnet trojan, ie. a plague affecting people's stuff - non-sentient accoutrements whose existence and existential value depends entirely on their utility to their associated human.
There's no clear consensus on which of these is a better analogy for VW avatars - the first and third are a lot more common than the second, though.
To be fair to them, they're talking about it because health researchers are starting to think about virtual worlds in the context of epidemiology, and whether they'd provide a better model for studying the spread of infectious diseases than standard computer models.
There are a lot of unexamined assumptions there, though, and it sounds as though this team really do need a virtual world designer or theoretician on board. The first big one that jumped out at me is that there was quite a lot of griefing going on, with many players deliberately infecting others. The pat answer to 'why is this a problem for the model?' is that this behaviour doesn't happen to nearly the same extent in real-life epidemics (I'm sure it happens in some cases), but there's a deeper issue as well.
Deliberately infecting others involves sacrificing your character's life - that doesn't mean much in WoW, but the important point is that it means very different things to different classes of player. (If you want to go into this in more detail, the reference is Richard Bartle, and his matrix of player types.) And I'm not at all sure that this level of heterogeneity, of differing levels of commitment to (investment in) the world, translates at all to the real world.
The second big one is that the ethics of the case are problematic to say the least.
[A] major constraint for epidemiologists studying disease dynamics at the moment was that they were limited to observational and retrospective studies.
For example, it would be unethical to release an infectious disease in real life in order to study what the consequences might be.
So, of course, it would be entirely ethical and uncontroversial to release an infectious disease in a virtual world, without explicit opt-in or any of the usual niceties. Really it would. And the other problem, of course, is getting a virtual world with a decent population to do it in. You certainly won't attract anything near the required numbers (ie. large enough so you can study them statistically with confidence, and do the same with subgroups - large enough so they can be thought of as a fluid rather than a granular, er, blob) just by wanting sign-ups for an academic study. You won't be able to do it in an existing large VW without i) cutting a deal with the designers, who will not unreasonably want to know what's in it for them, and ii) risking some serious legal problems. Different jurisdictions treat virtual property (and virtual people) very differently, and if a third party is putting them at risk then it's very possible that that third party could suffer some consequences. It's a near certainty that they'll suffer from the process, even if found in the clear.
There are basically three levels of analogy on which we can think about the digital Black Death experiment (which hasn't been done, and I suspect won't be). The first is the deliberate spread of a plague affecting humans. The second is the deliberate spread of a plague affecting people's farm animals or pets (foot & mouth disease, for instance), ie. non-sapient (but still sentient) dependents with their own concerns distinct from their human's. The third is a botnet trojan, ie. a plague affecting people's stuff - non-sentient accoutrements whose existence and existential value depends entirely on their utility to their associated human.
There's no clear consensus on which of these is a better analogy for VW avatars - the first and third are a lot more common than the second, though.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2007-08-21 11:29 am (UTC)no subject
Date: 2007-08-21 12:58 pm (UTC)Also, the model is not very real, to be honest. The main disease vector was the NPCs who wandered around infected and infecting but never being killed by the plague. These NPCs follow pre-designed routes through the city and never alter this path for any reason. Therefore the main reason the disease spread wasn't actually human at all. The reason that attempts to spread the disease on my server failed was because they either didn't manage to infect the NPCs or the NPCs had been changed to make them immune from the disease.
no subject
Date: 2007-08-21 12:59 pm (UTC)The Economist and the Financial Times were all over this story two years ago - doing what everyone else does, picking up their science news from New Scientist on a Friday morning. The BBC can't even do that any more.
Maybe they just trawl old magazines when one of their three thousand News & Current Affairs staffers visits the dentist on a slow news day.
BTW, I thought that some WoW players were deliberately infecting pets and smuggling them out as plague bearers. Count that as one 'miss' in the BBC coverage. Another is that both pets and characters have long-term value in WoW, so you get a very different set of behaviours to more abstract 'what if' simulations where the operators aren't really players with an emotional investment in the outcomes.
I have stated this before: if you're interested in futurology and really disruptive ideas, start watching for something totally left-field coming out of either virtual worlds or MMORPG's. The virtual plague and its consequences fizzled out two years ago, but the potential of these networks for real-world modelling has been conclusively proven... And no, I don't think that's the Next Big Thing either. Sometimes, it's best not to be too specific in your predictions.
(no subject)
From:no subject
Date: 2007-08-21 01:34 pm (UTC)