Preventing Lone Wolf Shootings with AI-Powered Investigation Tools

May 2024

Among the many criminal threats that face the world of the 21st century, mass shootings may be among the most concerning. These attacks are swift and unpredictable, most often targeting the most vulnerable of populations in schools, shopping centers, and entertainment venues. Unlike the activities of established terrorist organizations, lone wolf attacks usually involve an individual with no identifiable affiliation to violent groups, making that person’s plans and activities difficult to detect before they strike. As of the writing of this article, the United States has already experienced 16 school shootings in 2024, including the notorious attacks in Perry, Iowa and Arlington, Texas, and the thwarted attack in Mount Horeb, Wisconsin. Another mass shooting recently took place at a private pool party, one of three mass shootings in Florida in a span of three days.

The good news is that these attacks are no longer as unpredictable as they used to be due to the trail of clues that investigators can sometimes find through internet sources. It may surprise many people to learn that lone wolf mass shooters, from radicalized loners to bullied teens planning revenge, usually publicly post social media content that hints at their planned crimes well before they take action. In some cases, they might communicate directly with radicalized groups via social media to gain ideas and encouragement. In other cases, they communicate their personal feelings and desires to a more general social media audience. For instance, in the case of the recent attack in Indiana, the shooter made several posts regarding being bullied, as well as pictures of himself with firearms accompanied by songs about gun violence and ominous slogans suggesting that he was planning to take action.

In most reports about these tragic events, the relevant social media content is surfaced only as part of the investigation after the crime, a process to uncover motives, possible group affiliations, and potential related threats. While social media investigations are certainly valuable after the attack, sophisticated Artificial Intelligence (AI) tools can also enable investigators to engage in preventative measures which could stop mass shootings before they can happen.

There are many approaches to preventative AI investigations of potential mass shootings. If authorities have reason to suspect that a particular individual may be radicalized or otherwise dangerous, they can analyze open source (i.e., publicly available) information to search the worldwide web, deep web, and dark web for online materials linking the person to hate speech, extremist groups, known forums for expressions of violent thought, platforms for the purchase or modification of weapons, and a number of other types of significant information. Such an investigation can help investigators to learn if a threat exists, and how to proceed.

Investigators can also base their searches on particular places, events, and times. They can use AI tools to find publicly visible social media posts that match locations (e.g., schools, entertainment venues, stadiums) or events (e.g., games, parades, ceremonies) with online references to violent activities. The best AI investigation software systems can apply sophisticated lexicons and translation software to find and analyze coded speech, slang, emoji codes, and 100 different languages. As Pat Cheatwood, former Pelham, Alabama Chief of Police noted about his work to ensure large event safety, “The peace of mind that this technology offers is priceless — being able to monitor publicly accessible social media posts for threatening language 24 hours a day, without having to have an officer or analyst at a computer around the clock is a great way to give city and county leaders the peace of mind they need for their communities.“

Predictive analysis and prevention are among the most difficult challenges that analysts can face. When attempting to prevent tragedies as unpredictable as mass shootings, AI investigation tools appear to be our best hope.

Share this post