NORTHCOM’s 20-year-old data analytic processes
are undergoing a refresh.
US surveillance aircraft watch overhead as adversary forces move from one region to another, feeding imagery to a machine to analyze. Artificial intelligence (AI) sorts through the trove of data and quickly recognizes an abnormality in the number of vehicles moving through a parking lot. The AI queues both military and commercially available sensors to refocus on that location. Military leaders across the globe, and up the chain of command, are connected and able to see that information in real-time, providing the “decision space” needed to come up with deterrence options that hopefully limit the risk of accidental escalation.
That’s U.S. Northern Command boss Gen. Glen D. VanHerck’s vision for a new homeland defense design—and it’s not that far from reality.
“I’m a firm believer that [someone], in the future, with the right data, at the right time, will win, whether that be in crisis, … in conflict, or day-to-day competition,” VanHerck told Air Force Magazine.
Russia, and other countries, will absolutely take advantage of civil unrest in our country.
Gen. Glen VanHerck, USNORTHCOM commander
NORTHCOM was created 20 years ago in the wake of the 9/11 attacks on New York and Washington, D.C., and though the threats have evolved significantly since then, the command still largely relies on the same analog systems and processes to share threat data and intelligence. Solutions are mostly regionally focused and still involve human analysts entering data into spreadsheets, verbally providing updates across operations centers, and pulling together PowerPoint slides to brief leaders.
For example, during the U.S. Air Force-led operation to evacuate Afghan refugees to safe havens in August, the operations center had to stand down every four hours so analysts could “review disparate spreadsheets and ensure numbers matched,” according to an October white paper from the Center for Strategic and International Studies (CSIS).
Such a process might have been effective enough against a “low-tech, slow-moving adversary,” but it could prove “disastrous” in a conflict with a peer adversary like Russia or China, wrote Emily Harding, deputy director and senior fellow of the CSIS International Security Program, and Air Force Col. Matthew Strohmeyer, a military fellow at CSIS. Strohmeyer was the lead planner for the Air Force’s first two Advanced Battle Management System (ABMS) on-ramps, and he planned three-more similar experiments for U.S. Northern Command before coming to the Washington, D.C.-based think tank.
“While analysts are assembling data and attempting to communicate, an adversary could be in the late stages of conducting a cyberattack that severs communications with far-flung forces, preventing a response,” according to the paper “From Data to Insight: Making Sense out of Data Collected in the Gray Zone.”
“A complete data picture for early, effective warning will be critical,” they added.
VanHerck, who is dual-hatted as the head of NORTHCOM and the North American Aerospace Defense Command, agrees. That’s why he is pushing to modernize the United States’ communication and warning systems to give leaders time to better define options before launching a kinetic response to potential threats.
“I don’t believe that we can surround our country—or North America for that matter—with kinetic endgame defeat mechanisms,” VanHerck told Air Force Magazine. “It’s unrealistic and unaffordable.”
Evolving Threats
VanHerck said Russia remains the No. 1 threat to the homeland, because of capabilities it’s developed that fall below the nuclear threshold aimed at disrupting, delaying, and degrading the United States’ ability to project forces forward in a regional conflict.
As of the end of March, NORAD had conducted more than a dozen intercepts of Russian bombers and maritime patrol aircraft, an increase over previous years. The Russian aircraft sometimes loitered in or near the U.S. air defense identification zone for hours at a time. Russia also launched several disinformation campaigns in recent years, aimed at undermining U.S. democracy and breaking the will of the American people, VanHerck told Air Force Magazine.
“Russia, and other countries, will absolutely take advantage of civil unrest in our country,” he said. “On a day-to-day basis, they fanned the flames of COVID vaccination policies out there, through disinformation campaigns, to make us—internally [in] our own nation—believe or not believe what they want us to” believe.
Although China today poses a much bigger threat to U.S. forces and allies overseas, it is only about five to 10 years behind Russia in terms of being able to kinetically strike the homeland, and it is “on par” with Russia in terms of nonkinetic capabilities, VanHerck said.
According to the Defense Department’s annual report on China’s military power, released on Nov. 3, China is building nuclear weapons significantly faster than previously anticipated, it already has a “nascent nuclear triad,” and will field more than 1,000 nuclear warheads by 2030.
Rapid development of new stealth aircraft, the expansion of that aircraft’s weapons-carrying capacity, and the shift of China’s air and naval forces from a defensive posture to one of power-projection and long-range strike also is eroding the United States’ long-held advantages in the air domain, according to the report.
The Financial Times reported in October that China had tested a nuclear-capable weapon that allegedly circled the globe before reentering the atmosphere at hypersonic speeds. Although the missile reportedly missed its target by more than 20 miles, it’s clear China has made significant progress when it comes to hypersonic weapons.
Alexandra Baker, the Biden administration’s nominee to be deputy undersecretary of defense for policy, told senators shortly after the test “there is a sense of urgency” for the U.S. to develop similar capabilities.
“They have pursued a strategy of seeking to blunt U.S. advantages over a number of years, not only in terms of hypersonics, but also in space, counterspace, [and] cyber,” she said during her confirmation hearing.
Increasing the Decision Space
Although VanHerck recognizes there are some scenarios where a kinetic strike may be a necessary first response, it should not be the default option. He wants to use domain awareness, machine learning, and artificial intelligence to get into an adversary’s “cognitive space,” to better understand what they are thinking, and give senior U.S. leaders time to come up with options that enable the U.S. to gain the operational and strategic advantage.
When VanHerck assumed command of NORTHCOM in August 2020, he wanted to build on what the Air Force was doing in its ABMS experiments, which already had some success connecting sensors and data in ways not done before. But the Advanced Battle Management System was focused on a kinetic endgame, and while VanHerck says this is important for the services, he wanted his team to move left of the threat and look instead for ways to prevent the threat from getting to the homeland in the first place.
“The combatant command was feeling an urgent need to increase our homeland defense capability and to try to meet that need, because the regular capability development process just wasn’t developing it fast enough,” Strohmeyer told Air Force Magazine.
The series of NORTHCOM-led experiments that followed, called the Global Information Dominance Experiments, or GIDE, are focused on enabling a globally integrated deterrence. Though NORTHCOM does not use the phrase joint all-domain command and control (JADC2) when talking about GIDE, VanHerck acknowledged that’s what it is. JADC2, he said, can cover anything from the tactical to strategic level; his focus is on broadening the “decision space to create deterrence and de-escalation options, and if we must, defeat options.”
Instead of focusing just on detecting, tracking, and engaging a cruise missile, for example, Strohmeyer said GIDE considered the entire life cycle of that cruise missile threat. What is its pattern of life? When was it placed on the launch platform? What was the launch platform doing before launch? How can machine learning and artificial intelligence help the U.S. better understand the normal day-to-day pattern of life? How does that information help the U.S. better understand when a potential adversary is planning to act against us?
“Rather than us just having to react to a detected cruise missile over North America, instead we could have machines that are giving us alerts, saying, ‘Hey, I’ve seen in the last 24 hours a change in the pattern of life of the number of vehicles at this adversary bomber base. You might want to take a look at that,’” explained Strohmeyer, who planned the first three GIDE experiments. “Then a human can take that curated information, look at it and go, ‘Oh, that is interesting. There’s something going on here.’ And we can combine it with other sources of information or intelligence, and then ultimately go, ‘Yeah, they might be moving down the path of doing something that we don’t want them to do.’”
Armed with that knowledge, multiple combatant commands can simultaneously work together to come up with proactive options “to deter that adversary from taking that unwanted action,” Strohmeyer said.
GIDE 1, which took place in December 2020, was a tabletop exercise that involved four combatant commands. Participants took advantage of new technology that pored through historical satellite imagery and data from previous actions, and then alerted them to changes in an adversary’s pattern of life. However, the focus was really on breaking through existing stovepipes, integrating that data, and finding ways combatant commands can better work together.
GIDE 1 allowed multiple commands to operate in a common collaborative environment where everyone was seeing the same warnings from the artificial intelligence at the same time.
“In the past, you didn’t have combatant command collaboration happen until you’re in front of the Chairman [of the Joint Chiefs of Staff] or the Secretary [of Defense],” Strohmeyer said. “We had combatant command collaboration starting at the very lowest level.”
GIDE 1 was so successful, the second experiment conducted in March 2021 brought in all 11 unified combatant commands, and utilized live forces participating in an exercise occurring simultaneously in the Arctic.
GIDE 2 used live readiness data from the Air Force’s Project Brown Heron, which is an effort to combine previously stovepiped data sets together. So, instead of having to make a phone call to learn what aircraft are available to move as they plan sorties, the information was already available in a common cloud-based data system.
The overwhelming response from the CCMDs was, “Why aren’t we using this [operationally] now?” Strohmeyer said.
Changing the Culture
Speaking at a Center for Strategic and International Studies virtual event in August, VanHerck said, “The tools that we demonstrated are ready to be applied at the operational to strategic level to create time and decision space.” The challenge, he later told Air Force Magazine, is changing the regionally focused culture.
VanHerck often says that homeland defense doesn’t actually start in the homeland. It starts with allies, partners, and other combatant commands working together to deter and defend forward locations, so a strike never makes it to the homeland.
GIDE 3, which took place in July, once again included all the CCMDs, as well as interagency partners like the Joint Artificial Intelligence Center (JAIC), and it brought in several allies and partners to observe.
During the experiment, a GIDE team aggregated early indications and warnings from 120 days of geopolitical events, using real-world alerts to highlight adversary actions, once again requiring collaboration across combatant commands. Leaders then used artificial intelligence to review deterrence responses.
In the conflict stage, a military Blue Force faced off against Red Force threats at Michigan’s Alpena Combat Readiness Training Center, as a cloud-based computer network, accessible both in Michigan and at Tyndall Air Force Base, Fla., enabled participants to share data and demonstrate responses.
The focus of GIDE 3 was trying to understand how much earlier the U.S. can learn about competitor and adversary actions, Strohmeyer said. During the experiment, the JAIC also demonstrated its Matchmaker capability for “machine-enabled crisis deterrence and conflict defense,” according to a NORAD/NORTHCOM release at the time. Strohmeyer said instead of the old PowerPoint and Excel spreadsheet analysis, machines provided “real-time options,” though humans still made the final decisions. The team also partnered with the Defense Department’s Project Maven, which uses machine learning to identify people and objects in intelligence imagery.
“I think today, especially with Russia and China, they’re going to be global, all-domain problems, whether that be in day-to-day competition, or crisis, or conflict,” VanHerck said. “There’s no more regional problems, and so changing that culture first, I think, is important.”
He said while DOD’s strategies and operational plans are global and all domain, “the way we develop and design a force, budget, and acquire [capabilities, as well as] the way we train, should also be starting out with a global perspective.”
Real-World Application
Col. David Morgan, NORAD and NORTHCOM’s division chief for strategic analysis and experimentation, said the goal for GIDE 4 and beyond is to expand the involvement of allies and partners.
Though in the near term that likely will include Canada and NATO members, he said, “As the platforms develop, we are aggressively working on technical and policy issues to ensure our allies and partners can participate in future GIDE iterations.”
GIDE 4 is slated for early spring and will coincide with several already-planned U.S. and allied exercises, Morgan told Air Force Magazine via email.
Whereas the first three GIDE events primarily used AI and machine learning against real-world historical data gathered during “previous competitor activities,” Morgan said GIDE 4 will focus more on the operational level.
“In GIDE 4 we are already applying machine learning and AI tools in real-time toward recognizing changes in the pattern of life to provide better indications and warnings earlier than traditional means have provided in the past,” he said.
Through its work with Project Maven and the Joint Artificial Intelligence Center, GIDE continues to “challenge the traditional acquisition cycle,” he added, with the goal of pushing new technology to troops faster.
“GIDE has demonstrated the power of software-based technology and the rapid development that can be achieved in teamwork with leading industry partners,” he said. “Since GIDE 3, JAIC and Project Maven have developed new features and improved capabilities while working with users across the Department of Defense and industry. As planning continues, we expect GIDE 4 to provide a venue for demonstrating and experimenting with other related technology sponsored by the other combatant commands.”