LAUSR

Abstract In counterinsurgency, agent networks are double-edged swords. They are useful tools for degrading insurgent influence and protecting the population. However, they also endanger the population in some ways, as we have seen with mass executions of suspected agents and agent misdirection of raids. Identifying how/why this occurs is critical for developing intelligence practices to more effectively implement COIN strategies. This exploratory study uses three recent counterinsurgency cases – for which significant secondary open-source agent network documentation is available – to identify, describe, organize and analyse patterns of noncombatant-targeted violence associated with human intelligence networks. Identified cases of noncombatant-targeted violence in Operation Enduring Freedom – Afghanistan, Operation Iraqi Freedom, and the – Israeli–Palestinian conflict are used to develop a theoretical framework that models intelligence-related violence incentivization, which also draws from theoretical literature on human intelligence, civil war violence, police states, community policing, and the sociology of betrayal.