The Intelligence Everyone Can See
A new field of analysis that taps open source imagery is coming into view under the watchful eyes of remote sensing.
Commercial satellite companies are giving rise to a new space revolution, launching hundreds of small satellites into orbit to do what the U.S. military cannot or at least will not do: photograph practically every inch of the Earth every day. The result is an explosion of geo-enabled unclassified information that has turned the imagery-based discipline of geointelligence on its head.
This change could even produce a new breed of intelligence analyst that exploits imagery and geospatial data from the unprecedented fount of unclassified information.
Initially, some in the intelligence community were elated by the treasure-trove of open source solutions that systematically enhances intelligence gathering and processing, says Terry Dyess, team leader and operational planner of a U.S. Southern Command (SOUTHCOM) experiment that marries commercial capabilities with open geointelligence (GEOINT). But that euphoria was quickly subdued by concerns that images of entire swaths of land and oceans were becoming available to everyone—friend or foe, Dyess says. “While we were going, ‘Hot diggity dang, we’ll be able to see a lot of things,’ the awareness immediately became, ‘Whoa, wait a minute, they’re imaging all of our stuff too,’” he says. “It’s as if someone jerked the shower curtain down, and we’re all showering without that shower curtain now.”
The distinctive SOUTHCOM experiment to leverage commercial and open source solutions is called COGINT, or commercial plus geointelligence. What links the two is a plus sign, not an ampersand, Dyess points out, highlighting that the two camps are not mutually exclusive.
Already, commercial satellite images inform farmers of peak harvest times, apprise aid groups of refugee traffic in camps and educate insurance companies about shifts in urban environments to set appropriate premiums. For the intelligence community at large, these space-based eyes are showing that they can help narrow the military’s situational awareness aperture. When coupled with open source automation tools, COGINT’s efforts have alerted analysts to noteworthy changes, such as military ship movements to ports of interest, renewed activity on runways used by known drug smugglers, or the deforestation of areas known for illegal gold mining, explains Brig. Gen. Kate Leahy, USAR, who was director of SOUTHCOM’s J-2 Intelligence Directorate until August 7. “There is huge interest across the intel community in figuring out how we leverage machine learning and automated change detection to make our imagery analysts even more effective than they are today,” she says.
COGINT’s mission success hinges on making sense of and extracting actionable intelligence from heaps of otherwise immaterial data contained within countless images. To do that, the team relies on advanced software, artificial intelligence and automated change-detection technologies. It figures out how best to dovetail open source offerings with unmatched military satellite systems, Gen. Leahy states.
“We don’t know of another team like this,” says Scot Currie, director of the Mission Integration Office in the National Geospatial-Intelligence Agency’s (NGA) Source Directorate. “The team’s dedicated focus on unlocking the possibilities of not only the new small satellite imagery but also the explosion in geo-enabled publicly available information [PAI] is unique. We have to unlock the potential of [GEOINT] better and faster than our adversaries, then apply our exquisite capabilities to stay leaps and bounds ahead of them. The genie is out of the bottle. This capability will be a reality in the world.”
The NGA awarded a $20 million contract a year ago to Planet, which designs, builds and operates small satellites. The partnership provides the NGA with global imagery almost daily. Known as the Planet Feed, this service is being used across the National System for Geospatial Intelligence (NSG) and by select members of the civil federal community.
This space revolution has the trappings of the commercialization of the Internet and GPS, both of which started exclusively as Defense Department programs. “They were strategic advantages for the United States. Now they are available to the entire world—including our enemies,” Currie says. “That just happens when governments fund technological leaps. One day, those leaps become mature and the new normal. When that happens, the rest of the world benefits, and the governments that used to enjoy a strategic advantage have to find new advantages.”
The irony of an intelligence-based effort so reliant on open source and commercially provided products is not lost on the planners who work in an otherwise risk-averse community. The changes cause concern and uneasiness, Currie says. “That’s a lot of cultural transformation for our agency, and it’s about getting people comfortable with working in that area. These new data sources don’t easily fit into our classified analysts’ workflow, nor do they easily move through our current enterprise architecture,” he says.
Pointing to a proven model such as COGINT helps allay the apprehensiveness, Currie adds. “Having a dedicated team leveraging it outside of the ‘normal’ system is allowing us and SOUTHCOM to explore the real operational utility before the capabilities are mature enough to be scaled across the enterprise,” he says.
The COGINT team comprises traditional satellite intelligence analysts, data scientists, social media experts and linguists. The fusion of personnel and assorted new sources has proved the operational utility of this new GEOINT capability to make sense of the tidal wave of new information, says Maj. Mike Little, USA, deputy chief of the GEOINT Division of the Joint Intelligence Operations Center.
“When it comes to the capability of using open source data and commercial imagery, it’s not a question of whether our adversaries are going to use it, it’s how good are they at it—because we know they’re using it,” Maj. Little says. “We need to start using the tools, the model, the methods. Because if we don’t start doing that, we’re going to lose this race.”
Attaining superiority over adversaries is a formidable juggling act, Dyess offers. “The way to stay ahead is to do it better and faster than they do it on the unclassified side and then use our exquisite capabilities as the game changer,” he says.
To be clear, the new commercial imagery and geo-enabled PAI will not replace the need for either exquisite military-grade equipment or human analysts. “No one is expecting analysts to allow this new capability to replace current capabilities,” Currie surmises. “This is about using the new capabilities to allow traditional analysts to offload some of their current work while gaining the ability to get tipped that they need to conduct further classified analysis.”
“We will always need that really great human analyst to play this very human game of intelligence,” Dyess echoes.
The overall COGINT concept grew out of a much bigger NGA experiment called GEOINT Pathfinder 2, last year’s massive retraining program to increase the percentage of government data scientists and technical and coding talent. The idea was to reduce reliance on contractors and grow an in-house talent pool readily available to address changing needs, says Chris Rasmussen, the NGA’s public software development and GEOINT Pathfinder lead. “When you look at the aggressive small satellite launch schedule and what’s happening on the ground with everyone taking pictures of everything, you need new skills to stay afloat as GEOINT commercializes,” he says.
Growing skill sets within the government and the private sector is as important as growing trust among analysts, who place equal emphasis on data and algorithms, says Jane Chappell, vice president of global intelligence solutions for Raytheon Intelligence, Information and Services. “The more data you have, the better and richer the analytics are that come out of that data,” she says. “But they have to make sure that [analysts] understand the pedigree of that data, whether it’s commercial or open source. Where did that data come from? How accurate is that data? Have the pixels been manipulated in any way that would [affect] the intelligence?”
And they must trust the algorithms working behind the scenes that make sense of the data deluge, Chappell adds.
Intelligence analysts, in general, spend about 80 percent of their time looking for the data they need to analyze, she says. “If we can get the data to find the analyst versus the analyst finding the data, then we’re making the analyst much more effective right off the bat,” Chappell says.
COGINT’s efforts are picking up steam and piquing the interest of other intelligence community leaders, Currie says. But before they can effectively spread, leaders must tweak acquisition processes to let agencies buy commercial geospatial Earth observation data, products and services faster and easier. The NGA and the General Services Administration (GSA) formed the Commercial Initiative to Buy Operationally Responsive GEOINT, better known as CIBORG, to do just that.
Carefully vetted companies register products and services through the GSA, creating a go-to list of suppliers that agencies can easily tap. “The era of multiyear, billion-dollar contracts for services that last decades have had their time,” NGA Director Robert Cardillo said last year of CIBORG. “We have to become more agile and revisit fundamental acquisition strategies.”
Geointelligence is one of many topics that will be discussed at the Intelligence and National Security Summit September 6-7 in Washington, D.C.