In the face of extreme violence, after the initial shock and sadness, there is a renewed sense of urgency to understand why the attack occurred and to prevent it from occurring again, especially when the violence stems from a radical ideology. More and more often in the aftermath of these attacks, the internet is littered with breadcrumbs of comments, messages, and interactions of the perpetrator showing an escalation in thinking.

To name a few, the New Zealand shooter left behind a video, countless comments in online message boards, and a manifesto that references internet memes and the writings of other radical white supremacists involved in mass violence. The Isla Vista shooter in California also left behind a manifesto and video. His manifesto has influenced subsequent violent extremists connected to a radical wing of the incel (involuntary celibacy) sub-culture.

Given the online paper trail these individuals leave behind and the ideology that undergirds these attacks, prevention might start by considering how predictive technology could help us identify online “hot-spots” that embolden these kinds of attacks. A good preventative program would work to diffuse individuals in these online communities that espouse extreme positions and are poised to turn violent. Beyond preventing mass-scale attacks , prevention techniques could also focus on preventing suicide and self-harm among members in these groups.

This isn’t the first time predictive analytics have been used in crime. From a street crime perspective, cities have been experimenting with predictive software to attempt to guess and mediate crimes. Chicago has developed and used not just predictive crime map but a list of which Chicagoans are most likely to be involved in crime – either as a perpetrator or a victim. The city uses this list to both make arrests and investigate but also to provide social services – although some data shows that arrests outstrip offers of social services. Additionally, there has been concerns that Chicago’s and other cities’ predictive policing programs perpetuate racial assumption and previous patterns of policing.

Outside of a typical street crime, predictive policing has already been employed in an effort to curb international terrorism, both by recognizing patterns in location, timing, and other factors to locate “hot-spots” and aggregating lists of people based on a profile and other discrete characteristics. One of these predictive policing programs is called countering violent extremism (CVE) which gives grants to community organizations, U.S. attorneys’ offices and police departments to identify and dissuade individuals “at-risk” of terrorism.

However, these programs have been sharply criticized for their practice of flagging religious practices or other innocuous behavior as suspect and as having a disparate influence on Muslim communities as well as simply failing to consistently distinguish radicalized individuals. Nevertheless, even some critics think that with evidence-based criteria, these programs could have success.

These concerns about government surveillance continue to be real concerns as people have proposed to expand CVEs, not just to Islamic terrorism, but to groups organized around racism or misogyny. Considerations of these concerns is worthwhile. At the same time, given the reoccurrence of these racist or misogynist attacks, it would be prudent to examine ways in which big data and predictive models can help us derail and prevent violent extremism stemming from all kinds of radical political ideologies, not just radical jihad. Perhaps the correct course of action is not CVEs in their current form, but continued research in order to develop a evidence-based model to recognize patterns in online communication and other factors that precipitate these attacks.

— Rachel Miklaszewski

 

Comments are closed.