Entity Extraction Combats Terrorism Financing

Entity Extraction, Homeland Security, Intelligence Analysis, Risk Management

Entity Extraction to counter terrorism financing


How Terrorist Groups Exploit Charities

Charities that operate exclusively in the U.S. rarely have a problem with terrorist organizations or individuals exploiting them to promote their own goals. However, U.S. and more generally global charities that operate outside of their home country and provide funding to local charitable organizations have a significant challenge in dealing with terrorists seeking to misuse them.

That’s because terrorist supporters frequently infiltrate and use these local charities as a cover for raising and moving funds, persons, military supplies, and other resources. In addition, they will often use funds siphoned off from a local charity to provide social and other humanitarian aid to local vulnerable populations in order to gain more support from them.

In addition to these legitimate charities being exploited, there are completely fraudulent ones that are totally dedicated to terrorism-supporting activities.

How Financial Institutions Check Sanctions Lists to Find Bad Apple Charities

A critical juncture in ensuring that there’s no exposure for a charity to terrorism-supporting activity is when a U.S.-based charity asks a financial institution to transfer funds to an affiliated organization or individual abroad.

At this point, of course, the required action is for a bank to check the established sanctions lists (OFAC, etc.) to look for bad actors, both persons and organizations, as well as Politically Exposed Persons (PEP) who have been identified or suspected of funding terrorism.

Finding Adverse Information by Monitoring the World’s Media

There’s another source of data, though, that has previously proven difficult to exploit: adverse information about terrorist supporters or actors found in the world’s media.

Adverse information regarding charities would include media reports of their being investigated or penalized by authorities. Such information can appear in a variety of sources:

  • Traditional news media
  • Blogs and web articles
  • Social media and internet forums

One advantage of monitoring media is that it is up to date compared with sanction and PEP lists, which are typically a bit outdated.  Of course, it has been difficult for any organization, no matter how large, to monitor the world’s media because of its sheer volume. This is where Entity Extraction can help – it can analyze a large volume of news sources in real time for information about individuals or organizations that indicates terrorism involvement.

Entity Extraction, an AI Technology, Enables Adverse Media Monitoring

Entity Extraction, aka Named Entity Recognition or Extraction, is a groundbreaking A.I. technology that supports adverse media monitoring at scale. The news media and the other sources mentioned above contain very large amounts of unstructured text data, well beyond the capacity of purely human efforts to stay on top of it. In the search for terrorism-related information regarding potential customers which are charities, Entity Extraction has the capability to quickly process a wide range of media in order to collect this information.

Entity Extraction automatically identifies semantic concepts in text such as names of people and organizations. Advanced Entity Extraction goes beyond just named entities and identifies more complex semantic concepts such as events.

For example, Advanced Event Extraction captures those events where governments apply sanctions on corrupt charities. Here’s one example. It’s an excerpt from a news account describing the sanctioning of a Pakistani charity by the U.S. and Saudi governments:

“The United States and Saudi Arabia on Tuesday sanctioned a Pakistani charity for allegedly financing violent extremist groups in Afghanistan and Pakistan under the guise of humanitarian work. The sanctions target the Al-Furqan Foundation Welfare Trust.”

Advanced Event Extraction would produce structured, normalized data from this unstructured text specifying:

EVENT_TYPE: Sanction


o   United States

o   Saudi Arabia


o   Al-Furqan Foundation Welfare Trust

This structured data, when persisted in a database, would be easily searchable and analysts at a financial institution would be able to easily determine that this charity was corrupt.

In sum, Entity Extraction provides sophisticated means to identify terrorism-involved charity organizations with a vast volume of world media.