Leveraging Artificial Intelligence to Counter Terrorism

September 2021

Established and emerging terrorist groups show remarkable similarities in making use of open sources and online platforms to spread their poisonous messages, find and gain financial and other support, and cultivate new members. Whether they are religious extremist terrorist groups, far-right nationalist groups designated as “terrorist” by the US and other governments, or local violent insurgencies associating with terrorist groups for support, their transnational organization and communication are hallmarks of their operations. Their willingness and support for “lone wolf” attackers supporting their causes – and urging their supporters to use vehicles or knives for lethal strikes when mass-casualty weapons are not available – further spreads and heightens the threat most countries face from this new form of terrorism.

The use of open-source communications by terrorist groups has facilitated their fund-raising and recruitment capabilities but is also a double-edged sword: The “signal” these groups are leaving essentially “in the open” is demonstrably the source of valuable information for tailored technologies. The development and deployment of Artificial Intelligence (AI) to help the Subject Matter Experts, advanced analysts, and skilled investigators save time by running pre-set queries on these ever-growing datasets further deepens the ability to leverage technology to both establish a powerful and reliable “early warning system” for emerging general and specific threats, as well as to investigate individuals, groups, and organizations in-depth. These investigations can be triggered by the “signal” provided on a continuous basis by the AI, to prevent a threat from becoming a lethal operational act, or, for those attacks which could not be stopped, to provide a near real-time investigations capability to identify and arrest the perpetrators – and identify their networks.

While many “lone wolf” attackers are independent and unpredictable actors, they often still leave a notable, identifiable digital footprint as they develop into threats.  Many “lone wolf” terrorists use online platforms to share ideas, find a supportive audience, express their hatreds, and gain information that is crucial to their causes and missions. This is an additional avenue for AI to explore and consider as it combs through oceans of daily-generated publicly available data that would take teams of experts and experienced analysts, countless hours to process and analyze.

This powerful AI capability is not limited to open sources: Even Dark Web postings and communications can be very effectively and productively analyzed by AI solutions. Some AI programs have been remarkably successful in identifying purchase patterns on websites, with parameters that include both the types of items and quantities that fall into ranges commonly associated with terrorist actions. These often include weapons, ammunition, weapon accessories, travel, chemicals that can be used in explosives, and even commodities associated with human trafficking (sometimes used for terrorist funding).

One of the critical capabilities AI delivers is the pattern recognition and entity extraction/resolution that allows for signals to be amplified for analysts and investigators to check and look into further.

These high-impact AI solutions are effective in preventing terrorist activities when the information they provide is acted on promptly and ethically.  There is a range of standards and regulations on the critical issue of bias, and the importance of “explaining” the reasons for flagging people or behaviors as meriting a closer look or follow-up actions.

Share this post