Proceed with caution: on the use of computational linguistics in threat assessment

Large-scale linguistic analyses are increasingly applied to the study of extremism, terrorism, and other threats of violence. At the same time, practitioners working in the field of counterterrorism and security are confronted with large-scale linguistic data, and may benefit from computational methods. This article highlights the challenges and opportunities associated with applying computational linguistics in the domain of threat assessment. Four current issues are identified, namely (1) the data problem, (2) the utopia of predicting violence, (3) the base rate fallacy, and (4) the danger of closed-sourced tools. These challenges are translated into a checklist of questions that should be asked by policymakers and practitioners who (intend to) make use of tools that leverage computational linguistics for threat assessment. The ‘VISOR-P’ checklist can be used to evaluate such tools through their Validity, Indicators, Scientific Quality, Openness, Relevance and Performance. Finally, some suggestions are outlined for the furtherance of the computational linguistic threat assessment field.