Ghosts in the Machine
February 23, 2008 by wwallaceIs your tin-foil hat affixed? Yes? Good.
National Intelligence wants to know what you’re doing in virtual worlds, and to detect suspicious patterns in your behavior. There’s a very informative unclassified report from the Office of the Director of National Intelligence to Congress discussing in brief their objectives and reasons for this system.
The interesting section: they propose a program called “Reynard” (which Wired points out was the name of a trickster figure in medieval Europe) which is “a seedling effort to study the emerging phenomenon of social (particularly terrorist) dynamics in virtual worlds and large-scale online games”. It goes on to propose that the project studies social norms in virtual worlds and online games, and thus learns how to look for suspicious behavior amongst players. The program is a pilot, and may “increase its scope to a full project”.
Couple this with the paragraph before this in the report, which discusses the “ProActive Intelligence” project, which examines “causal relationships that are indicative of nefarious activity”.
The number of potential issues this raises is immense, but there are a couple that come to mind that are particularly ridiculous, but not completely impossible.
Narc-Bots
This is going to sound silly, but hey, lets take the government document at its word and take it to an extreme.
You thought gold-farmers were bad. Government terror-detection algorithms are loosed on bots in popular game worlds—uberspy007 on Habbo ? Totally a narc-bot—that level 1 guy in the Barrens who has been sitting in the same place all day? Yeah, narc-bot. The bots look for chat patterns, key words like “bomb”, “beheading”, “Iraq”, log references per character, and if they flag a user as a potentially interesting subject, they request his or her IP information from the developers, and take their investigation out of the virtual world.
If developers were cooperative, they wouldn’t even need bots—just chat logs, game data, or relevant metrics. Or hell, subpoena the information! Do wiretapping laws carry over to virtual worlds? How do you admit evidence gathered in a virtual space into a real life courtroom, or explain it to a jury? There’s a number of interesting precedents that could be set that would pave the way for further legislation and regulation of online communities.
Distributed Intelligence
The power of distributed computing is often touted as a way of achieving super-computer like results for cancer research, finding alien signals amongst radio telescope data, and a host of other altruistic ideals. Distributed intelligence, then, would be using a network of civilian informants in virtual worlds to gather information, report suspicious behaviors, and help gather or supplement the raw data necessary to put together the pattern matching that IARPA/DHS obviously wants. Homebrew intelligence analysts—the same kinds of people who track spy satellites from their backyards for fun—run loggers and data-mining programs while they go about their daily grind online that pump data back to government analysts. It’s like planting a wire on a turncoat mobster, except that there’s no danger of being shot when your “buddies” discover the Feds are listening.
I don’t see either of these things happening soon, but the important takeaway is that in a Congressional report looking at various terror prevention projects, a project to study burgeoning terror threats in “virtual worlds and large-scale online games” was one of the five major bullet points. Participation has become high enough that the spy agencies are interested in monitoring virtual worlds—and as unregulated as they are, I’m sure its a goldmine of grey areas and unchartered domestic spy-craft territory.


