When science challenges terrorism
Can science become central to combating terrorism and organised crime? Predictive policing uses mathematical and analytical tools to identify potentially dangerous people and forecast crimes. In 2011, Time magazine even named it as one of the year’s best innovations. These predictions normally draw on data on past criminal acts (considering type, location, time, etc.) with statistical methods, to model “hotspots” that become the focus of specific prevention policies. An example is the PredPol software used by police in many US cities, and created eight years ago by researchers from the University of California (UCLA) and Santa Clara University. However, there are significant challenges in risk-based measurement when it comes to forecasting human behaviour because of its unpredictable nature. The neutrality of data collected may be biased by the assumption that crime will likely happen where it has already occurred in the past and committed by the same people. The risk is that the algorithm could unfairly target specific groups, such as minorities. Moreover, predictive programmes present several “vulnerabilities” when it comes to counter-terrorism, as underlined by Andrew Ferguson, professor of Law at the University of the District of Columbia, Washington: “Crimes happen every day, several times a day, and with certain crimes occurring in repeated patterns in particular geographic locations. Terrorism instead happens rarely and even more rarely repeats in the same place. Whatever lessons you have learned about risk from street crime will not be applicable to predict the risk of international terrorism.” Researchers are trying to take a step forward, focusing on the affiliation process of criminal minds. They are analysing the formation, radicalisation and dissolution of organised crime and terrorist networks, integrating for the first time social, psychological and economic data. Through Agent-Based Modelling (ABM), a class of computational models, the scientists are currently developing a virtual society, to test different scenarios. Their aim is to go beyond predictive policing techniques, providing governments with the opportunity to evaluate, in the long term rather than the short, the impact of potential policies on the recruitment and radicalisation phases. The study comes under the EU project Proton. One of the members of the consortium, Maria Laura Fiorina, jurist from University of Pavia, Italy, explains: “These models will address questions such as: if the policy maker increases the number of schools in a specific area with high potential of criminal recruitment, what will the impact be on this?” “Data heterogeneity is a fundamental aspect. A low level of instruction is proven to be a breeding ground for organised crime affiliation, but the same principle does not apply to the terrorism networking process. This is an example of the challenges of selecting the elements that will be introduced in the ABMs.” Fiorina warns: “Our research is aimed at helping governments to identify risk factors. However, it is also fundamental to elaborate a correct assessment of the impacts, to evaluate whether the prevention measures could cause social stigmatisation towards group already suffering from discrimination.” “In this context, a study realised under Proton on the London Muslim community showed that prevention policies, such as the stop and search method had damaging effects, worsening the situation. This approach may make some people more radical in their thoughts and actions,” Fiorina adds. Read more: https://www.projectproton.eu/science-challenges-terrorism/
Keywords
crime, terrorism, security, predictive policing, statistics, modelling
Countries
Belgium, Italy, United Kingdom, United States