Social Media Intelligence versus the War on Lone Wolf Terror

In a post 9-11 world, the war of terror rages on, billions of dollars are being spent to defend and destroy those that would cause harm to US citizens, her allies, and interests. In recent years, it seems the threat that may cause the greatest harm to our security is not longer Al Qaeda, but potentially those that already reside within our borders. These individuals may have succumbed to foreign propaganda or environmental pressures which inspire them to turn against their fellow citizens in an outward violent act. Unlike an attack by Al Qaeda, a lone wolf or wolves are near impossible to detect by traditional intelligence networks and local law enforcement. Moreover, the lone wolf terrorists that are already domiciled within the borders of the United States have ready access to lethal materials and high population centers. Yet, in a world highly connected via social media, mobile communications, and the internet, could social media be used to sniff out potential terrorists threats? Personally, I believe it could.

As US Citizens, we need to think on a much more localized level, while simultaneously monitoring global conditions. The challenge for law enforcement is that Lone Wolf terrorists may be inspired by global events without actually being directly connected to a specific terror organization or watched intelligence network. In a statement by former FBI Agent Jack Cloonan, “We don't know what a lone wolf; al Qaeda-inspired operative looks like. We don't know where they hang out; we don't know really what motivates them. So when you don't know that, you've got a talent pool of people that is so huge, it stresses law enforcement. We just don't know what they look like and what they want to do."

Here inlays the problem. The US government currently looks at social media intelligence in the same contextual fashion as a Google search.  Most recent example is its 2010 contract with Visible Technologies via the CIA technology investment fund, whereas Visible crawls over 500,000 social content sites twice a day, and scores the poster sentiment (positive, negative, neutral) against a predefined contextual document. If the poster has a degree of influence or social following, then they may end up on a watch list by one or more government agencies. This is the same type of technology that many companies use to measure their brand reputation. Where this technology fall short is it sorts more on contextual relevance than actual meaning. Moreover, influence over others means very little if you are a lone wolf.

A better platform would be one that uses contextual search like visible in association with semantic search and scoring.   Whereas, the process of semantic search seeks to improve search accuracy by understanding searcher intent and the contextual meaning of terms as they appear in the searchable data space, whether on the Web or within a closed system, which generates more relevant results. Example: Tom Cruise, being the bomb, would not be confused with a man named Tom looking to detonate an explosive device on a cruise. More importantly, this tool should be available to both local and federal law enforcement agencies. Given that if local law enforcement is reporting suspicious activities IE acts of trespassing at city water reserves, and there seems to be a repetitive actions happening in other cities within a given time period then there may be cause for a deeper federal investigation. If this was compounded by lone wolf or wolves social media chatter on “Let’s poison the water supply”, using action words and phrases showing a repetitive intent to comment such and act then the treat matrix would be elevated to mobilize against such a threat while in the planning stages.

How would one build such a system? I spent a lot of time on this issue, given that I have designed several social analytics and engagement tools in the past. Simplistically speaking it would require the following steps.

1)      Access to large data pipes of RSS feeds with specialized crawlers designed to target specific sites of interest. Include video and media feeds as well.

2)      Create model documents of potential terrorist scenarios

3)      Turn the model  document into a searchable algorithms using Semantic Latent Indexing

4)      Build information capture sites that use the model documents as a guide to search engine optimize a lone wolf into accessing “How to build bad things that hurt people” downloads.  

5)      Run the search algorithm to find specific “suspects” of interest per potential terrorist scenario.

6)      Use an automated reverse lookup or trace route (if possible) to determine origin and potentially identity of user.

7)      If user is anonymous, then auto engage the user, using a social bot, or automated response program that uses in part the suspect’s subject matter in conjunction with a preformatted action text in hopes to gain more intelligence on the user to establish true intent, by posing as a like minded individual. Engagement tactic may be to send them to the capture site, capture their IP address, time on site, downloads, to help establish potential threat.  

8)      Compare data gained in suspect search algorithm with that gained from the capture site.

9)      Find affiliated accounts, handles, or suspect’s network of friends via communication threads or social media friends and begin to semantically monitor conversations.

10)   Monitor and turn potential treats over to local, federal, or national law enforcement based upon treat analysis.

11)   Threats that became creditable would be tagged in the system to better target future threats. The end result would be a fairly accurate profile per threat scenario.   

Outside of Al Qaeda terrorists, this type of technology would have a great practical application in local law enforcement in regards to:

1)      Potential sex offenders / human traffickers

2)      Pre-planned. mass violence – such as the 2007 Virginia Tech Shooting leaving 33 dead

3)      Drug trafficking

4)      Organized crime  

I would love to hear your thoughts. Please comment and share. If anyone knows someone in government that would want to build such a system, I would be happy to have that discussion in depth, so please contact me.