Physicists Create Algorithm That Maybe Be Able To Predict Terror Attacks

Clear and Present Danger

A team of physicists has just created an algorithm that studies the behavior of pro-ISIS online groups and predicts their future actions, including when they may become real-world threats.

The team focused on a Russia-based social platform called VKontakte, and identified 196 pro-ISIS online groups, singled out due to the group's propagation of actual ISIS content, pledge of support for ISIS, or calling of jihad in the name of ISIS.

The study, published in Science, found that the number of these online groups—which offer a form of anonymity and freedom of speech for individuals—dramatically increase in the lead-up to real-world events. This allowed the group to create an algorithm that leverages this information to predict when things will occur.

Surprising Solutions

Eliminating terrorist activity on social media presents a challenge—often shutdowns come from the platform itself, which must navigate the line between public safety and free speech. But even with efforts from platforms and governments and even hacktivist groups—Anonymous removed 20,000 Twitter accounts with ties to ISIS last year— the challenge still remains

This is because, as hackers or online moderators shut down groups in sites like VKontakte, the members split into, or join, smaller groups. Then these eventually coalesce into larger groups, which get shut down, and so on.

While the algorithm promises to be no silver bullet to ISIS in social media, the study does recommend on how the problem should be approached.

"The math shows there’s a very precise rate at which aggregates [online groups] need to get shut down. If you don’t want the pro-ISIS or pro-extreme aggregates to grow into one huge global aggregate, then you’ve got to shut it down faster than this precise formula that we’ve come up with.", says lead author Neil Johnson in an interview with Pacific Standard.