By PATRICK STEPHENSON
BRUSSELS – Lone actor extremism — more commonly known as “lone wolf” terrorism — is a growing threat in Europe. On 14 July 2016, a lone actor drove a truck into crowds celebrating Bastille Day in Nice, killing 86 people and injuring 434. Every few months a suspected extremist attacks police or transit officers in Belgium, for example, usually with a knife or axe. Police and civil defence authorities desperately hope for a method to prevent such attacks, or at least mitigate their consequences as they occur.
Finding such a method is the goal of PRIME (PReventing, Interdicting, and Mitigating Extremist events). A three-year research project with a EUR 3.6 million budget, PRIME launched on 1 May 2014 and is focused on the different phases of a lone actor threat.
Six main objectives define the project’s work: describing the risks poses by lone actor extremist events (LAEEs); mapping key factors and processes at all stages of the event; translating the resulting risk analysis into a ‘meta-script’ of LAEEs; identifying moments of possible intervention, or ‘pinch points’; delivering requirements for the design of measures for the prevention, interdiction, and mitigation of LAEEs; and developing communication measures that can either help prevent or mitigate attacks.
SECURITY EUROPE previously reported on the project in its earlier phase, but it is now approaching its conclusion in April. For a sneak peak, we spoke to its project coordinator Dr Noemie Bouhana, senior lecturer in security and crime science at University College London.
The first problem, she explained, is a simple dearth of data about lone wolf incidents.
“These are events that almost never happen,” she said.
While the variables that seem to cause such attacks — ideology, social isolation, unemployment — affect millions of people, only a few will ever become attackers, making it difficult to sort out signals from noise in the data. Furthermore, the data that PRIME initially possessed on lone actor attacks covered only North America and Europe, and the researchers hoped to make conclusions that were broadly applicable in wide variety of countries, cultures and situations.
“We’ve looked at all the cases, and it’s barely more than 100. So for me, as a criminologist, there’s not that much information,” she observed.
PRIME got a break when its researchers partnered with experts at Hebrew University in Jerusalem. Its academics gained access to data from ‘run-over’ attacks — when vehicles are used — and stabbing attacks that had never been interpreted before. This Israeli dataset was as large as the US and European datasets put together, and it greatly broadened the scope of their analysis.
The project categorised attacks into three distinct phases. The first, explored by team members based at Aarhus University in Denmark, was radicalisation and how someone comes to see terrorism as something acceptable, and even necessary. Researchers at Leiden University in the Netherlands tackled the second phase, attack planning. This examined how a belief in terrorism becomes action. It also includes how a target is selected, how it is prepared for, and how an actor maintains his or her motivation. Researchers at Hebrew University focused on the third phase, the attack itself.
The project also included other teams. Partners at the University of Warsaw examined the gap between what data reveals and what practitioners can mistakenly believe. Other colleagues at King’s College in London focused on communications. For example, when prevention and detection have failed and the attack is in progress, how do you communicate with the public to mitigate the attack? Conversely, how can the public communicate with security services to let them know what’s going on?
PRIME also brought in computer scientists to help model the data using different graphical approaches, including Bayesian statistics and probabilities. “The main advantage of this approach is that when you get more info — say, for 2016 — you can put in this info in the network and it will recalculate all the probabilities without having to start from scratch,” Bouhana said.
Among its deliverables, one of the project’s most important will be its theoretical framework for lone attacker behaviour which should help future researchers determine what questions to ask, before an attack occurs. “I believe you only get the right answer if you ask the right question,” she said. “This [framework] is the thing that people will neglect. They just want stats. But a major contribution will be this framework.”
Bouhana hesitated to reveal the full preview of PRIME’s conclusions, but she did allow some glimpses.
“Offender profiling has never worked,” she said. Instead, where the radicalisation occurs geographically may reveal as much about future attacks as who gets radicalised. “As the environment where the radicalising settings are found changes, then the sort of people who find themselves in those settings changes. In theory, if you had good intelligence on where the radicalising agents are, then you would have some basis on which to make some predictions or guesses on what the next generation of radicalised people will look like,” she said.
Another finding — not entirely new to students of extremism — is that the phrase ‘lone actor’ is a misnomer. “Throughout the process, they have social contact,” Bouhana said. “Even social isolation is, in fact, a social process. So thinking of them as people in their basement that no one has heard of, seems to be incorrect. Which is, in fact, good news.”
PRIME will present its findings at a final conference scheduled at meeting of the Royal Society on 7 April in London.
That is the dilemma behind ‘profiling’: the omnipresent false positive. Bouhana, to her credit, was clear that PRIME cannot offer, and does not pretend to offer, a magical formula that will pick out signals from noise. What she does offer is a framework that can be used to lower the number of lone actor events – in part, through a better focus on where, geographically, radicalisation occurs – whether in Brussels’ notorious Molenbeek district, or online. In an age of outsized claims about the benefits of ‘behaviour detection’, such humility is refreshing. We look forward to reading PRIME’s conclusions next month.