Predicting how terrorists behave using “Big Data” and AI

24

September

2018

No ratings yet.

After the tragedy of 9/11, everybody was astonished that airplanes were used as a weapon for the terroristic attack, for that had never happened before. In 1994 however, the writer Tom Clancy published a book in which an airplane was used to attack the US government. This led to an enormous brainstorm session in Washington, attended by filmmakers and writers, to come up with possible scenarios for another attack.

This inspired former film maker Peter de Kock, who got his PhD in 2014 on the dissertation “Anticipating criminal behaviour”. This dissertation provides the building blocks of a computer model that predicts (the narratives of) terrorist attacks. The model is based on an enormous database of terrorist attacks that happened, but also on terrorist attacks that were created on a creative basis, in for instance a book or a movie. According to the model, every attack has a narrative, and the narrative consists of building blocks, the so-called anatomy of the attack. The narrative entails a wide range of scenario components, examples include: a protagonist, an antagonist, an arena, a modus operandi, symbolism and a red herring (a false clue). In total there are twelve components, and with these components, every story can be told. All the attacks in the database, whether they were real or fiction, are dissected into these twelve components. Each component has its own sub-components (e.g. bomb or rope), in total adding up to more than 200.00 sub-components. The compilation of all this data is mostly done by text mining.

To let the model run, as much components (by extracting them from a situation) as can be found are being put into the algorithms of the models. Then, the model compares this with the hundreds of thousands of narratives and sub-components that are in the database, looking for a correlation between some of the narratives. Resulting in scenarios where a plate of copper found in the east of the Netherlands can be traced back to a Jihad-warrior who used a similar plate for an attack in Syria and who was from Dutch descent and recently returned to the Netherlands. Since a couple of years AI is used to create new scenarios based on the data in the model. The model can thus be compared to a chess computer, thinking a few steps ahead and analysing every potential move a terrorist can make. In this way governments and companies can protect themselves better against potential threats, after all humans are only unpredictable to a certain extent.

Sources:
https://www.pandoraintelligence.com/index/

Click to access De_Kock_Anticipating_10_09_2014_emb_tot_10_12_2014.pdf

Please rate this

Leave a Reply

Your email address will not be published. Required fields are marked *