New Tools to Create Time and Information: “Building the Bike While We Ride It”
Two things I can never have enough of as a commander are information and the time to consider it. Unfortunately, today’s strategic landscape offers me less of both than I’ve had in the past. For much of the past three decades, the U.S. military was able to project power across the globe without worrying about a conventional attack on its homeland. This level of security was unique and it allowed the United States to address threats while they remained far away from U.S. shores. However, America’s adversaries have not been standing still. The country now faces competitors capable of striking discrete military targets or critical infrastructure in the United States using cyber weapons, hypersonic missiles, or other limited conventional means of attack. This makes the country less safe, limits the options available to senior U.S. leaders in a crisis, and could rapidly degrade the military’s ability to surge forces from the homeland into other theaters.
As the commander responsible for the defense of North America, I believe that the military ignores these threats to the homeland at its peril. Peer competitors’ doctrine, operations, public pronouncements, and demonstrated exercises all show that they view the threat or conducting of conventional attacks on the U.S. homeland as part of a viable strategic concept to defend their interests. Should such an attack occur, the United States would be forced to choose between three unfavorable options: launching a preemptive offensive strike that risks tripping the nuclear threshold, accepting an adversary strike on the homeland and then subsequently retaliating (again with the risk of nuclear escalation), or de-escalating on terms favorable to the adversary.
The United States needs more time and better options to deter conventional threats to the homeland. The key to developing these options is data. As America’s adversaries become increasingly capable of attacking U.S. territory, the military will need rapid access to information and intelligence, predictive analysis tools to identify trends and potential adversary actions as or before they occur, and communication networks that speed up the pace of decision-making.
With more time and information, commanders can posture forces ahead of competitors’ actions, strategically message them through diplomatic channels, or align their commands’ actions with allies and partners, ultimately creating doubt in the mind of adversaries that they could achieve their objectives.
A New Homeland Defense Design
North American Aerospace Defense Command (NORAD) and United States Northern Command (USNORTHCOM) are focused on a new homeland defense design that leverages information and data to deter and defend against global threats. Specifically, the two commands are trying to improve how information is collected, analyzed, and shared through a series of Global Information Dominance Experiments. With origins tracing back to 2019, the series represents my efforts as the commander of NORAD and USNORTHCOM to realize the Joint Staff and Defense Department’s concept of global integration. It also provides a means to shift the department away from regionally focused plans, strategies, force management, and force design to a paradigm where combatant commanders instead plan and coordinate across combatant command boundaries. Potential adversaries aren’t bound by the lines that the Department of Defense has drawn, and will compete and fight wherever the most gains can be achieved.
Since conventional strikes on the U.S. homeland are likely to be part of China and Russia’s strategy to counter American involvement in a crisis, the Defense Department needs to adopt a global perspective for homeland defense. This perspective should include participation from every combatant commander and the contributions of allies and partners, because actions to deter and defend the homeland cannot start in the homeland.
The United States does not have the resources to ring North America with a comprehensive network to defend against every single threat imaginable. As a result, NORAD and USNORTHCOM are working to identify geographic concentrations of key critical infrastructure in North America that need to be defended against traditional threats such as ballistic or cruise missiles, or against new threats such as hypersonic missiles, cyber attacks, or unmanned aerial systems. It includes the defense industrial base and civilian infrastructure sectors such as energy, water, and transportation. The recent example of the Colonial pipeline cyber attack shows just how important it is to encompass this broader definition of homeland defense.
Global Integration Through Information Dominance
Fundamentally, information dominance is about possessing the capability and tools to create tactically and strategically relevant linkages between information collected from sensors and intelligence sources in a way that affords decision-makers the opportunity to use that information to make timely decisions. NORAD and USNORTHCOM’s approach to creating information dominance began by framing the challenge not as a military problem, but instead as a data problem.
In many military systems much of the collected data is left on the cutting room floor. It’s never analyzed because of the gargantuan amounts flowing into the system. For example, radar data analyzed by NORAD and USNORTHCOM’s Pathfinder program, one of the constituent tools used in the global information dominance series of experiments, revealed that only 2 percent of information collected by homeland defense radars was displayed for operator analysis and decision.
By viewing the problem as a data problem rather than one of equipment or hardware, NORAD and USNORTHCOM were able to quickly improve capabilities and unlock terabytes of hidden information. Platform-based capability development efforts often take decades. But software development and cloud computing move faster. Unlike hardware procurement programs, software development is not industrially intensive, it takes advantage of Moore’s Law and the exponential growth of computing power, and it is uniquely suited to defense acquisition tools designed to spur innovation rather than traditional contract mechanisms used in hardware development. Finally, software development is also less expensive. The cost of the global information dominance series of experiments and the tools developed for it can be measured in the tens of millions of dollars thus far, a fraction of the amount invested in most major weapons systems.
Through the series of global information dominance experiments, NORAD and USNORTHCOM stepped outside the normal procurement cycle, and employed a non-traditional acquisition, testing, and fielding approach. As a result, we used the tools contained within the series even as the tools themselves were refined. In coordination with several Department of Defense innovation groups such as the Joint Artificial Intelligence Center, the under secretary of defense for intelligence and security’s Project Maven, the Air Force’s Chief Architect’s Office, the Defense Innovation Unit, and others, NORAD and USNORTHCOM and partners used small business innovative research contracts, cooperative research and development agreements, and the discrete application of existing contracts to move quickly from concept to experimentation. The effort, which began two years ago as a small initiative to look at how NORAD operators could better utilize existing radar data, has now expanded to a series of experiments incorporating participation from every combatant command. The experiments utilize sensor data from land, maritime, air, space, cyber, and intelligence sources, publicly available information, and new software developed to facilitate cross-combatant command collaboration, planning, force allocation, and decision-making. In short, we are building the bike while we ride it.
By ingesting data from multiple sources, the experiments are breaking down existing data stovepipes from legacy systems developed to look at singular threats and instead incorporating information into an open architecture that analyzes data from multiple sources simultaneously. For example, the experiment has brought in data from homeland defense radars, the maritime automatic information system, commercial and satellite imagery, electronics intelligence, readiness and maintenance, blue force tracking, and many other sources. This creates a data repository that allows for rapid aggregation, processing, analysis, dissemination, and display of data in order to identify trends with the goal of providing autonomous machine-developed cues of competitor actions hours or days before traditional methods of intelligence assessment. These cues could be indications of force movements, pre-positioning of weapons and logistics assets, or intelligence collection. The cues are then presented to human operators to confirm tactical or strategic significance.
The second iteration of the global information dominance series was conducted this spring and included participation from all 11 combatant commands in a wargame against peer competitors using both historic and current data. The goals of the experiment were to gain a common understanding and picture of the threat from each of the combatant commanders; to obtain earlier indications of adversary actions through AI and machine-learning tools; and to conduct cross-combatant command coordination on response actions, thereby obtaining faster decisions and ultimately improving deterrence.
The experiment was successful on all three counts. Combatant commanders’ staffs demonstrated a vision of what military operations in the future could look like. Beyond that, the experiment enabled those participating to think in new ways and develop processes to break down the regionally focused mindsets that hamper the Department of Defense’s coordination in competition and challenge the U.S. military’s ability to deter globally.
For example, during the experiment, a build-up of aircraft at a competitor’s base was detected, using historical data of exactly such a build-up to cue the systems’ AI algorithms. In today’s environment, dependent largely upon an analyst’s eyes and experience, such a build-up could be missed. Even if detected it would typically generate a response from a single combatant commander, since a coordinated response uses cumbersome manual coordination tools. Within the experiment, however, because each combatant commander received simultaneous notification of the indicators, representatives from multiple commands nominated response actions to generate a coordinated global response adjudicated by the lead commander. The feedback from participants was overwhelmingly positive, and my fellow combatant commanders and I were uniform in our endorsement for fielding these capabilities as soon as possible and for working on implementation in concert with experimentation.
Admittedly, there is some amount of risk in moving fast. Some avenues of research and experimentation may prove irrelevant or duplicative, leading to an expenditure of resources without obvious gain. But risk aversion cannot be the approach to risk mitigation. The benefits to moving fast, including increased knowledge and capability development for use today, outweigh the expenditures and any false starts that may occur.
Privacy advocates have also rightly asked how privacy concerns are being addressed. In short, every input, before being incorporated into the experiment’s tools, was assessed by the source agency and legal experts (particularly in the case of intelligence sources and publicly available information) to ensure that its use was consistent with U.S. law.
Following the second experiment, NORAD and USNORTHCOM took data available during a recent world crisis and processed it through the tools developed for the series. In doing so, we determined that had the global information dominance tools been in use before the crisis, actionable information would have been identified two to three days in advance of traditional methodologies. To guard against confirmation or selection bias as much as possible, in that the outcome of the crisis was known, the data ingested into the tools was not selectively chosen. Rather, available information was broadly ingested and analyzed through the tools, which produced machine-generated observations for human decision, in effect connecting disparate dots faster than human analysts.
The advantage of having more time to make a decision cannot be overstated. In a future crisis, it could mean the difference between peaceful resolution and unintentional escalation. With extra time, commanders could select deterrence options proactively rather than responding to provocation in a way that increases the risk of strategic deterrence failure. Commanders could begin mobilization or targeted actions to put forces in place to counter adversary movements, increase readiness postures, or use diplomacy to defuse the situation. The goal of all is identical: conventional deterrence options that show adversaries the risks of their actions exceed the gains.
These cross-combatant command initiatives to create information dominance are critical to address conventional military threats to North America. Through the series of experiments, we are taking an approach focused on producing the elements most important to any decision maker: time and options. By integrating more information from a global network of sensors and sources, using the power of AI and machine-learning techniques to identify the important trends within the data, and making both current and predictive information available to commanders, NORAD and USNORTHCOM are creating time to make decisions and delivering the opportunity for senior leaders to choose better options.
The way the challenge is being solved is also important. Software-based solutions are allowing rapid and iterative development at a more affordable cost. Finally, the series of experiments, by actively including participation from every combatant command, is promoting adoption of the global perspective needed to compete in today’s strategic environment. The next experiment in the series, which will occur in July 2021, will build upon the framework and success of the two previous experiments with an even more complex strategic competition scenario using real-world live data. The experiment will also showcase how the software tools designed for cross-combatant command collaboration, assessment, and decision-making can be used to enable global logistics coordination in addition to intelligence sharing and operations planning.
The next step is to transfer the leadership for these efforts to an appropriate entity within the Department of Defense that will move beyond the combatant commanders and those that see the need, to one that can more effectively orchestrate and direct change across the entirety of the defense enterprise. With these efforts, the Department of Defense will close the gap between nuclear and conventional deterrence options and reduce the risk of strategic deterrence failure. The future credibility of America’s deterrent hinges on whether or not this promising start can be turned into an enduring reality.
The views expressed are those of the author and do not reflect the official position of the U.S. Department of Defense.