http://www.wired.com/dangerroom/2009/04/pentagon-wants-to-replicate-anthros/ By Sharon Weinberger Given all the problems the Pentagon has faced recruiting anthropologists to work with the military, it may have finally hit upon an easier solution: create computerized virtual anthropologists to replace flesh and blood human beings. In a new request for research proposals aimed at small businesses, the Pentagon says it wants technology that would “enable accurate forecasting of a given populations’ potential responses to military relevant events…” While not outright suggested as a replacement for the controversy-plagued Human Terrain System, which sends anthropologists into the field with the military, the new request for proposals does say this tool would be “used to facilitate or to replicate wholly or in part many of the tasks that a human anthropological consultation would provide such as, counter-insurgency, reconstruction or support operations, allowing faster and more accurate development of social-cultural behaviors.” As a concept, this is hardly brand new. The Pentagon for the past few years has been pursuing the Human, Social, and Cultural Behavior Modeling (HSCB) program (and this proposal is likely part of that larger program). That effort looks precisely at these type of models and computer tools. Proponents of the work argue that such models are less about actual crystal ball-type predictions of “target populations” and more about providing training and decisionmaking aides, but they also have difficulty explaining how such technologies would eventually be used. This proposal suggests, for example, it will be used to actually predict local response to a potential military action: “The software will be user-friendly and intuitive so that commander level military members.” That last sentence is a little hard to picture. Such work has attracted criticism even from some anthropologists working with the Pentagon, who fear that attempts to “widgetize” social science offers an unrealistic approach to cultural training, and preys upon the military’s love of technology. (And of course, there’s also that eerie similarity between these modeling programs and Isaac Asmimov’s concept of psychohistory, which didn’t work very well even in fiction.) More intriguing about this proposal, however, is the reasoning for why virtual anthros may be better than the real thing: “Today in DoD, this analysis is conducted by anthropological experts, known to carry their own bias, which often leads to faulty recommendations and inaccurate behavioral forecasting.” --

※ 发信站: 批踢踢实业坊(ptt.cc) ◆ From: 140.112.217.6