Tanith originally posted this in a comment, but this warrants its own thread. Wake up, all
ye virtual-worlders.
Here comes the sinister-sounding Reynard Project, an effort by the U.S. intelligence community to detect violent extremists infiltrating online games. The eventual goal is to be able to “automtically detect suspicious behaviour and actions in the virtual world.”Maybe these guys are new to the Internet?
There’s some good discussion of this over at Terra Nova. You can even look at the original report with which the Director of National Intelligence commissioned Reynard (if you’re a real keener). Mike Sellers on Terra Nova takes a similar tack to Tanith (“Haven’t these guys heard of the Internet?”):
(bordering on bizarrely), the report says that
“The cultural and behavioral norms of virtual worlds and gaming are generally unstudied. Therefore, Reynard will seek to identify the emerging social, behavioral and cultural norms in virtual worlds and gaming environments. The project would then apply the lessons learned to determine the feasibility of automatically detecting suspicious behavior and actions in the virtual world.” [emphasis added]
I have to admit I’m unclear as to what might constitute “suspicious behavior” in virtual worlds, much less behavior indicative of terrorist motivations — would that be killing as many Night Elves as possible in a battlegrounds session in WoW, emerging as a successful and highly profitable mole in EVE Online, or just using pose balls to role-play anatomically unlikely scenarios with someone using a hedgehog avatar in Second Life?
What do y’all think? It’s a lot easier to track someone in a virtual world than a real one – and we’re all going to be spending more time in VWs in the future . . .
Filed under: internet Tagged: | espionage, intelligence, Reynard, surveillance, Terra Nova, Virtual Worlds
Again, I get the feeling that the people who’ve conceived of this idea don’t fully appreciate how very differently people often behave when they are online versus when they are interacting face-to-face. The anonymity of the Internet makes people brave enough to say and do things they wouldn’t if there were any consequences for their actions. What many people would consider suspect behaviour in the real world is often overlooked or brushed off by online communities.
While I think that there are circumstances where greater consequences for online actions would be good, the Reynard Project brings up a number of legal questions – can I legally be tracked by a U.S. agency if I’m accessing the Internet from Canada or England or China or Botswana? How do you determine who has jurisdiction – is it a matter of where I am (sitting in my living room in Montreal) or “where” I’m engaging in “suspicious” behaviour (on a World of Warcraft server that is physically located in the United States)?
I’m also concerned that if a project like this goes through, that it will create a Big Brother Is Watching mentality, and that people will feel less free to engage and exchange and expand online, which I think would be a shame.
You can say that again. Edward Castronova, who blogs over at Terra Nova, has been writing about the ‘Magic Bubble’ that keeps games fun, and the effects of bursting it.
For instance, you’re a corp. owner in EVE. You’re sitting on hangars full of goods worth real money. Then the Guiding Hand Social Club comes along and takes you down, stealing $16,000 worth of minerals and space ship parts.
Apart from the technical problems Tanith has lain out above, I’d say another important question is should you have legal recourse?
You can reasonably claim that you’ve had a lot of real value stolen from you. But on the other hand, part of the reason we play games like EVE is that we can rob, blast, gank and teabag our opponents all we want. That’s what makes the games fun, and what makes them challenging is that we’re up agains real people. Forget CONCORD – how much fun would EVE be if the real cops showed up every time you podded someone?
I think that a strong case can be made that this kind of policing in most virtual worlds would kill them. There should be some sort of ‘here-be-dragons, caveat emptor’ clause in the EULA you tick every time you enter the VW. When you go into EVE Online or WoW or whatever, it is in the knowledge that ye may be ganked and robbed and ye shall not sue.
Mind, that applies only to combat-heavy virtual worlds. In places like Second Life, it’s a whole different ball game. People might actually welcome real legal safeguards on their property there. But given that there’s no overt means of taking it from someone (you can’t kill someone if there’s no combat) it’s less of a point.
Which, in my mind, makes all this more sinister, not less. Because Reynard isn’t about someone stopping you when you do something dangerous or illegal. It’s about someone (more like some computer program) watching you all the time, automatically, building up a profile of your behaviour, and somehow determining if you’re ‘deviant’ or not.
Remember?:
Sinister, indeed, indeed. It’s so generic, so just very very slightly not innocuous, so easy to overlook . . . I’ve been researching surveillance a lot lately (big issue where I live at the moment) and there’s one phrase that keeps popping into my mind: The Banality of Evil . . .
What about if you’re one of a large number of players swindled out of $100,000 in in-game money and assets?
But that’s besides the point. The real issue is this:
This is why Reynard is going to be a disaster. Who sets the threshold that the software uses to decide when one’s behaviour has gone from guarded to suspicious? What parameters are used to determine that the group of players meeting to discuss their infiltration op present a threat to national security?
One of the reasons that people spend time in Virtual Worlds (and spend the money for the privilege of being there) is that there is an implicit understanding that participants have far greater freedom than they do in the real world. Perfectly nice people can carry out merciless acts of industrial espionage and sabotage, secure in the knowledge that their identities are protected and that they won’t be assaulted by some furious victim of their online shenanigans. For many “villainous” players, their online personas are just that – roles that they play when online. Is it fair to flag someone for pretending a terrorist mastermind?
The Reynard Project isn’t just some random dude shrieking into a mic about how you stole his magic sword. It’s the authorities observing and judging your online actions, and that’s bothersome.
Ok, ok, hang on a second. I agree that this is a potentially serious invasion of privacy. I agree that the border-less world of the internet is ripe for intel-gathering on a large scale. I’m still wondering, however, just what means exactly they plan on using to do this.
their objectives:
1. Video Analysis and Content Extraction – software to automatically identify faces, events and objects in video.
question: this might be good for youtube… but we can barely do this for high quality cctv cameras, and many systems still require copious human operators. So where is the processing power and bandwidth to actually do this going to come from? Echelon has been having problems like this for years. More modern systems must have the same problem:
example, quote: “codenamed “Baseball” is the Information Awareness Prototype System, the core architecture to integrated all the (now dead?) Total Information Awareness Program’s information extraction, analysis, and dissemination tools. Work on this project is conducted by SAIC through its Hicks & Associates, consulting arm that is run by former Defense and military officials and which had originally been awarded US$19 million IAO contract to build the prototype system in late 2002.” can you really build DEADALUS for 20 mil?
2. Tangram – A system that wants to create surveillance and threat warning system that evaluates known threats and finds unknown threats to issue warnings ahead of an attack
Great. I want a system that creates gold brinks, evaluates property values, and builds me one shiny new house in every state. Does anyone know how in God’s name they plan on accomplishing this monumental task?
3. Knowledge Discovery and Dissemination - This tool is reminiscent of the supposedly-defunct Total Information Awareness program. It seeks to access disparate databases to find patterns of known bad behavior. The program plans to work with domestic law enforcement and Homeland Security.
Bad behavior? As everyone has seemed to agree, bad behavior on the intertr0n is poorly understood and probably not at all correlated with terrorism, at least not in the traditional “islamist fanatics crashing things into buildings” sense. Wired’s ironic intro line is appropriate. The USA’s real-world intelligence gathering has enough problems; I don’t see how an automated system can do what they want it to with our current level of electronic intelligence.
What they really need are superbly trained specialists. Say, the TCTF or Section 9…
Rusty: Tangram, VACE and KDD are all seperate projects from Reynard, about which very few details are given in the report (yeah, I read it. Train rides are boring, alright?). The first three projects are enormously ambitious and have numerous logistical and developmental problems that you’ve very succinctly pointed out.
But how hard would it really be to create an algorithm that tracks what players are doing in Virtual Worlds and keeps an eye out for patterns? I’m thinking not very, because games like WoW already use in-house data mining to track in-game trends and make loads of information available to the public.
I’m wondering who will be writing the algorithms, what they’ll be looking for and if it’s even possible for them to cut through all the crap that goes on in Virtual Worlds to find actual nefarious activity.
Yeah, probably the people they’ve hired to make it. If anyone can assemble the talent and expertise to do something like this, the US military/Intelligence apparatus is a pretty good contender.
DARPA invented the Internet in the 60′s, don’t forget. British engineers invented Radar and the jet plane. Don’t underestimate teh POWER of the mil/int side.
I thought the Luftwaffe invented the jet plane.
Or did they just militarize jet engines?
Hmm, Wiki says that Englishmen did a bunch of the work on the engines, and then some sneaky Jerries turned around and built a plane a couple of years before the British finished theirs.
I’ve gone off-topic, but I regreat nothing; I just learned something. My brains am bigger!
That is probably true. On the other hand, the playing field was a lot smaller back then, although I suppose the quality of American military programming has increased at about the same rate the net has.
Still, I think this requires a creative solution. Like scanning virally or fooling internet communities to do the coding/monitoring for you.
No way. Softwarez is teh anwsers!
Almost a week after Objective 514 told you about Project Reynard, the BBC brings you the story . . . Objective 514 pwn!
screw you! I was about to link that!
Curse my clanking metal body!
[...] play networked games, for instance on Xbox live – and if so, are their game actions monitored in a project-Reynard-like way? ‘Cause if you’ve got a convicted killer, on good behaviour, cavorting around [...]