In 2011, the simple exploitation of an existing data set could have prevented a near disaster in northern Afghanistan.
Then, an entire operations center watched as the feed from an MQ-1 drone, newly reassigned from its original mission, displayed a growing group of protesters at the perimeter of a small U.S. forward operating base. Although conventional signals intelligence indicated a possible disturbance, full-motion video confirmed the severity of the threat only well after it had matured. Intelligence analysts didn’t understand what the protestors were doing — and why they were doing it — until they had already massed at the entry point. If used properly, automated social media monitoring and geofencing, which calls for creating virtual geographic boundaries, could have filled this critical gap in situational awareness.
Goodbye gates, guards and guns
New technologies are transforming physical security from “gates, guards and guns” to an imperative that’s increasingly reliant on data systems. Efforts like last year’s pilot program between the Air Force and AT&T to establish a perimeter network utilizing multi-protocol label switching and SIM chip technology via an LTE network at Maxwell Air Force Base are the new normal. That system created a wireless smart perimeter with infrared sensors and facial recognition to detect and identify intruders as well as to alert base personnel of potential security breaches.
The military’s use of biometric scanning through facial recognition software, license plate “grabbers,” UAV and aerostat surveillance, and radar and seismic detection sensors are nearly requisite for access control. The application of technology of the internet of things — as evidenced by nearly $9 billion of federal money in 2015 — plus an embrace of artificial intelligence feed into the growing chorus for “smart bases.”
Harnessing networks for geo-situational awareness
Yet, one application of new technology could quickly exploit existing data sets and be rapidly employed, likely preventing surprises like the one in Afghanistan in 2011. Using geofencing to surface high-quality, publicly available information, layered atop conventional active and passive physical security measures, would enhance situational awareness around military bases.
While many bases already exploit social media data and integrate big data streams into a common operational picture, like Northrop Grumman’s Critical Incident Response System and AT&T’s Common Operational Portal, the U.S. military can more widely apply and emphasize this concept overseas. For starters, analysts at the tactical level could determine the latitude, longitude and radius of the area of interest around a forward operating base and adjust this geofence for monitoring. Designated social media aggregators would then comb available application program interfaces to surface information within the bounded area. These aggregators would deliver real-time data (e.g., geotagged photos posted to a variety of social media platforms) for analysts to triage and then paint a comprehensive intel picture, similar to targeters teasing out patterns of life through a watered-down version of activity-based intelligence. Such methodologies to identify patterns and predictive indicators — often through automation — are already staples of private sector and U.S. law enforcement physical security practices.
To go further, the use of natural language processing for sentiment analysis within this open-source intelligence would improve indications and warning. Sentiment indicators that identify potential threats in multiple languages are being used in the private sector and by U.S. law enforcement, with obvious applications overseas. Additional advantages of this big-data approach are a prodigious and archivable metadata trail, and the potential for interagency information-sharing that leverages the link analysis proficiency resident in many government organizations.
Potential obstacles to implementation
Despite its quick deployment potential overseas, domestic applications for the U.S. military’s use of geofencing and social media monitoring are limited. In a stark example of litigation risk, Facebook, Instagram and Twitter suspended social media monitoring startup Geofeedia’s access after a challenge from the American Civil Liberties Union in 2016. Renewed reluctance by social media companies to share data with third parties, borne out of reported misuse of user data by the firm Cambridge Analytica, would likely hinder future development of similar technologies. And until costs are offset by more widely applied AI technology, monitoring and assessing the veracity of ingested information will eat into analytical capacity. Yet, the responsible integration of open-source methodologies that reduce risk and improve situational awareness surrounding U.S. military bases is worth the roadblocks. Otherwise, the risk of surprise is too great.
Kara Frederick is a research associate for the technology and national security program at the Center for a New American Security. She has worked for as an intelligence analyst for Facebook and the Department of Defense.