A serious game empowering the prediction of potential terrorist actions

Law Enforcement Agencies (LEAs) are nowadays taking advantage of a wide range of information and intelligence sources (e.g. human intelligence (HUMINT), open source intelligence (OSINT), image analysis (IMINT)) in their efforts to anticipate potential terrorist actions. However, activity prediction based on these sources requires datasets for training reasoning algorithms, which are not typically available. In this paper we introduce a gaming approach to the generation of datasets that can be used for training reasoning algorithms that are able to predict the likelihood of terrorist actions against specific assets and locations in urban environment. At the heart of our approach resides a twoplayer game, whose moves map to actions that are usually associated with the various stages of planning, preparing and executing terrorist attacks in urban environments. Apart from presenting the game, the paper introduces also the range of reasoning algorithms that have been trained based on its output. It also explains the terrorist semantics that underpin the implementation of the game and the design/modelling of the rules of the game. Early validation results demonstrate that the presented reasoning algorithms can successfully classify terrorist activity, through distinguishing it from seemingly suspicious but unrelated events that are typically generated by non-terrorist activities.