Full Program »
Opponent Modeling of the Taliban in Afghanistan's Emerging Peace Talks
Terrorist groups can be modeled as rational actors that adapt the structure of their networks to perform reward maximizing actions in a given environment. In the current talk-and-fight environment of Afghanistan, the Taliban need to adapt their network to two domains, negotiations and combat. Opponent modeling can be used to understand how the Taliban concurrently optimize the structure of their network across both domains. Conceptually, it is first shown that negotiations have a different reward distribution – which yields a different optimal structure of the Taliban network – than combat. Empirically, it is validated that relevant changes in the structure of the network of the Taliban can be inferred from 2017-2018 data on Taliban attacks and Coalition airstrikes. Two alternative cases of a best response to these structural changes are provided that enable focused diplomatic engagement of the Taliban in the emerging Afghan peace talks.