Using Games To Predict Terrorist Actions?

Thanks to Popular Science for its feature article discussing the opportunities for using game-based simulations to predict the actions of "allies, enemies and even terrorists." The article explains: "The need for sim Qaeda agents is taking modelers down strange paths. The team at Moves [creator of the America's Army recruiting game] is trying to model the behavior and thinking of terrorists by creating a series of computer characters to populate a model code-named Iago, after Shakespeare's arch villain." However, Will Wright, creator of The Sims, injects a note of caution with regard to the general concept, pointing out: "As you scale up to larger and larger systems, you can probably model large trends... But what the Iraqi resistance will do over the next month is based on thousands of tiny local factors that seem to always be in flux and to be too granular to be modeled."

Comments

Didn't We See This Already?

[fuzzybunny]

...in the form of a geopolitical futures market?

It strikes me as interesting, trying to predict the actions of homicidal fanatics through whatever mechanism, be it something like this, which is essentially an extended human brainstorming, or through methodical, automated risk analysis.

It could be interesting to bring completely unrelated individuals' ideas into play, to see what someone pretty random with violence on their mind might go for (I mean, I'm sure that _someone_ uninformed would have come up with the idea of ramming planes into an office tower) but I would always be aware of a few caveats:

- Most people who would play something like this think differently (basic cultural mindset and all that) than a mad bomber from a middle eastern country (or Iowa for that matter)
- There's a danger of coming up with a lot of purely hypothetical red herrings--as anyone who works in security and who has ever held a brainstorming session to determine potential avenues of attack can tell you (no, it's not realistic that the Martians will attack with death ray spaceships, although it is conceivable)
- The more factors are considered in putting together a "risk" big picture (such as having a ton of online gamer geeks come up with ideas to blow up as many civilians as possible in one go), the greater is the human propensity to see said big picture as "authoritative".

That said, if it's just used as a tool to model potential avenues of attack, not a bad way of going about it.

Re:Didn't We See This Already?

[Ath]

I disagree a bit about this. I think it is wishful thinking to conclude that someone who follows a "normal" life and does not have homicidal tendencies could think of unpredictable ways to initiate terrorist acts.

For example, I believe it was a commission in 2000 (chaired by, I think, Paul Bremer) that warned of the use of airplanes as missiles to attack targets inside the USA.

Will Wright's point about granular details being too hard to model is only valid to a certain point and depends on what you are trying to achieve. If you are trying to model a world that predicts terrorist behavior, it seems we are far away from such abilities. If you are trying to model a world that gives a lot of freedom and capabilities to individuals, I think the real individuals will eventually demonstrate new and creative ways to do things.

I see a real danger here in that those individuals who can think in such creative ways would be branded as "high risk" individuals. There are a lot of people who assume you are dangerous merely based on your ability to think of violent behavior. The next step in this approach is that if you are willing to engage in such violent behavior in a virtual setting, you are more inclined to engage in the same behavior in the real world.

People have their own conclusions about whether any of that is true. Personally, the idea that virtual behavior or mind thoughts demonstrate a higher likelihood of such behavior is a little tough to accept. I can personally say that I have thoughts, fantasies and ideas that I would never initiate in the real world for a variety of reasons. I would never accept that I am a high risk individual because I think of such things. Of course some crimes are inevitably preceded by fanatasies about it. For example, rapists probably play out rape fantasies in their head before committing actual rapes. But going the other direction and establishing causation that all people who fanatasize about rape are potential rapists (or even inevitable rapists) is a bit ridiculous by most common sense measures.

The other risk is that such a modeled environment would also allow individuals who truly do have a propensity towards such behavior to practice their acts before engaging in them, even perfecting their plans. It is one thing to accept the benefits of pilot schools even when they are abused into training centers for terrorists. It is quite another to intentionally set up an environment where wannabe terrorists are given the tools for practicing terrorism, even if the goal is to prevent the behavior in real life.

Perhaps we could have a questionnaire at the beginning. Question 1: Are you a terrorist? Only those who answer no are allowed in the game.

Re:Didn't We See This Already?

[fuzzybunny]

Disagree? You're agreeing with me! That was part of my whole argument :)

To clarify; what I meant to say was that it's dangerous to rely on information generated by this sort of a mechanism as in any way complete or authoritative, as people who'd be involved in it (slashdot geeks?) probably have their brains wired different ways from some Pakistani religious student seeking an express ticket to see his 70 doe-eyed virgins and take as many of us corrupt, decadent infidel scum as possible with him in the process.

Rather, it'd be interesting to see what comes out of it, as no more than one of many forms of input or analysis. Hence my use of the word "brainstorming". The idea being that you will almost certainly not cover all the bases, but some novel approaches will inevitably result when you have a whole bunch of people not coming from (a) a terrorist or (b) an intelligence or government background thinking, "hmm, how would _I_ commit terrorist acts?" as a purely academic process.

Whether or not this would come up with anything really useful, whether one could separate the really feasible ideas from the chaff, and how much credibility one should assign to such ideas is left as an exercise for the reader.

Psico-history?

[BRSloth]

That reminds me the excellent "Fundation" series by Isaac Asimov. In that book, there was a man that created a new research line called "psico-history". That research allow him to say what would happen in the future by using statistics over a large group of people, and the predictions only worked on large groups, not on individuals. Someting like "a group will always work in some ways, but an individual will work randomly".

It seems that we are seeing the born of psico-history, using games.

Simulations have their limitations.

[no longer myself]

It's the same thing with hurricane predictions. Oh sure, they've got a decent grip on about where it's headed, but those things can change direction and completely nullify any predictions made 24 hours earlier.

Reality doesn't scale down to the simulation level very well when you need an extreme degree of accuracy.

"Terrorists only have to be lucky once. We have to be lucky every time." -can't remember who said it.

now i understand

[buttahead]

populate a model code-named Iago, after Shakespeare's arch villain

in my fonts, `I' looks like `l'. I thought lago would be a very slow implementation of logo. I like the shakespeare's villain idea much better.