Protecting Our GEOINT Advantage

During a keynote address at the GEOINT 2017 Symposium in San Antonio, former National Geospatial-Intelligence Agency (NGA) Director Robert Cardillo observed the 73rd anniversary of D-Day. Codenamed “Operation Overlord,” the indelible mission instantly turned the tide of World War II in favor of Allied forces.

“Operation Overlord was fueled by some of the most important maps and charts and imagery intelligence in history,” said Cardillo, who called D-Day a “pointed reminder of who we are and who we must be as an Intelligence Community.”

Two years later—on the eve of D-Day’s 75th anniversary—the Honorable Kari A. Bingen, principal deputy under secretary of defense for intelligence, shared the same sentiment during a GEOINT 2019 keynote address delivered on the same stage, in the same city.

“Much is known about the scattering of the paratroopers who assaulted Normandy,” Bingen said. “What is less identifiable in this success is the amazing accuracy of what would become geospatial intelligence that was provided to them prior to the invasion. … The precursor of GEOINT, 75 years ago today, directly enabled the success of the Allied landings in Normandy.”

The Department of Defense (DoD) has the same “solemn obligation” to warfighters today that it had to warfighters then, Bingen said during her Wednesday morning keynote. “Just as the imagery intelligence provided a critical advantage to the Allies that consequential night,” she continued, “so, too, does GEOINT provide us a comparable advantage today—so long as we continue to push the envelope in our design and employment of innovative GEOINT technologies.”

In an enterprise as entrenched as the Pentagon, pushing the envelope is easier said than done. Still, defense intelligence professionals must set ambitious goals that advance their tradecraft toward new horizons that will make future D-Day moments possible.

Sharing those goals with the GEOINT Community was Bingen’s charge on stage, where she outlined five principal objectives for the defense intelligence enterprise.

Goal No. 1, Bingen said, is to “move up the data ‘value chain,’ beyond raw pixels and bulk image buys, to services and applications that add value, knowledge, and insights.”

Doing so will require increased acceptance and integration of commercial capabilities.

“While NGA has vast GEOINT data stores, the commercial sector also has sizable commercial imagery libraries that are growing at an exponential rate,” Bingen continued. “With all of this unclassified data readily available, we are seeing an influx of new thinking, spurring new companies and business areas to achieve incredible innovation. They’re applying machine learning, advanced analytics, computer visualization, and offering new data products. We in government need to take advantage of the tremendous investment the private sector has made in this area.”

Goal No. 2, Bingen said, is fielding artificial intelligence (AI) and machine learning (ML), progress, which is evident in the evolution of Project Maven, established in 2017 by Lt. Gen. John N.T. “Jack” Shanahan, director of defense intelligence for warfighter support with the Office of the Under Secretary of Defense for Intelligence.

“I am pleased to report that in under two years, Maven has made steady progress as a pathfinder for AI/ML,” Bingen announced. “Most importantly, in under six months from the Authority to Proceed in June 2017, they delivered a real, operational capability to the field applying AI/ML to [UAV] full-motion video. Since then, the team has significantly expanded the capability—in scale and scope—expanding the number of geographic locations, number of platforms and data, and intelligence mission areas beyond video.”

Maven has delivered not only capabilities, but also lessons learned.

“What has been most significant is what we’ve learned about the ‘AI pipeline,’” reflected Bingen, who said Maven has been a classroom in which to develop scalable best practices in areas such as data accessibility, preparation, and labeling, not to mention algorithm verification, validation, and management. “Maven is now fielded in enough places that we’re hitting critical mass, and the next true measure of success will be to see actual changes in our analytical workflows.”

Those workflows complement goal No. 3: changing the tasking, processing, exploitation, and dissemination (TPED) model for GEOINT.

“We need to focus on the ‘ground,’ which has long played second fiddle to collection platforms,” said Bingen, emphasizing the need for domestic and international data sharing and interoperability. “We need a new, more dynamic, integrated model for TPED if we are to keep pace with the changing threat.”

Also needed are enhanced security and diverse talent—goals No. 4 and 5, respectively.

Of security, Bingen said it “is no longer optional, but a necessity” and called on industry to do a better job protecting trade secrets from foes such as China. “What’s in R&D now—what’s being stolen now—is what we’ll face on the battlefield in five to 10 years,” she explained. “And when we no longer have that technical advantage, our military advantage … erodes.”

Of talent, she said, “Our defense intelligence mission demands a diverse workforce. After all, our job is to understand others: other countries, their cultures, geographies, capabilities, and ways of thinking. So, diversity in backgrounds, ethnicity, gender, and in other areas across our ranks and in our thinking is vitally important to our mission success.”
The defense intelligence enterprise faces great challenges; but also, Bingen concluded, great opportunities.

“It is an exciting time to be in GEOINT. This Symposium is a testament to that. … Never has the government had so many diverse options available to it for GEOINT infrastructure, data, services, and applications,” she said. “We are at a true inflection point. We need to tackle our big data challenge; modernize our manual, labor-intensive processes; and better posture for intel in a contested environment.”