Decision Advantage in Competition Published Aug. 5, 2022 By Capts. Alexander Wright, Seamus Feeley, Shelby Copenhaver, Zachary Munoz, Timothy Bjorgan, and Jacob Garret (USAF) If decision advantage in conflict is closing our kill chains, then decision advantage in competition is keeping adversary kill chains open. In the 1950s, Col John Boyd embodied the concept of a decision cycle in conflict with his theory of the “OODA Loop.” He believed that the orient phase was key in gaining dominance over an opponent. In 2001, Lt Col Shanahan matured this concept in his National War College paper, “Shock-Based Operations,” in which he advocated a focus on the opponent’s orient phase. He believed that we should follow through on our own OODA loop in a manner and speed that breaks down the adversary’s decision cycle at the orient phase. This could induce an operational paralysis that would “prevent our adversaries from adapting to their ever-changing surroundings and cripple their ability to react to U.S. or coalition actions.” These concepts were developed to enhance decision advantage in conflict. Now in 2021, as we consider conflict less of an “on/off” switch and more of a spectrum, we will advance this concept one step further to help guide our efforts to purpose-build ABMS for delivering decision advantage in competition. Translating to the Kill Chain As we’ve now moved past the tenure of the OODA Loop, we can analyze the value presented by Col Boyd and Lt Col Shanahan in terms more closely related to ABMS: the kill chain. If we identify an analog for Col Boyd's key phase of opening the orient cycle within our construct of F2T2EA, the first three actions, F2T, are the best match. Subsequently, we can attach that logic to our enemy’s decision cycle. When an adversary opens their kill chain to begin operating toward a strategic goal, they enter the F2T phases which are equivalent to them entering the orient phase, per Col Boyd’s logic. If we execute our F2T phases faster and in a means that allows us to move to Target before the enemy can recognize it, we have a high chance of disrupting the continuation of their kill chain as they reconsider their actions. This effectively demonstrates a kill chain version of Lt Col Shanahan’s “Shock-Based Operations.” Proving Value in Competition With the goal of purpose building ABMS to attain decision advantage in competition, and the core component of this dynamic being how we carry ourselves in the F2T phases, we have scoped our analysis for guiding ABMS efforts to the F2T portions of its construct. As we press further, we need to confirm that our chosen scope will produce benefits in competition. When we consider F2T2EA, we usually do so in the frame of conflict since we typically think of ourselves operating on the kill chain after conflict has begun. But, if we examine the operations of the DCGS enterprise in the 480th ISRW, for example, F2T is performed every day in AORs where conflict is not yet underway. Specifically, DGS-4 performs and improves their F2T tradecraft for both USAFE and AFAFRICA without a move to the latter phases of the kill chain. We can draw from this that the F2T phases of the kill chain are regularly performed below the level of armed conflict and exists across the competition continuum (see Figure 1). Figure 1. F2T in Competition and Conflict If F2T are steps performed in competition, then they can also be considered our levers for attaining decision advantage in competition. If we want ABMS to deliver the decision advantage we seek then, per this logic so far, we need to focus on the F2T sensing grid. Doing so empowers us to follow through on today’s “Shock-Based Operations” in an environment that lives primarily beneath the level of armed conflict. Reorient and Maximizing for the Wins of Now With both our scope and relevance identified for how we want to guide our ABMS building efforts, we’ll want to look at where the most immediate connection points are for building out the sensing grid with the purpose of disrupting the enemy’s kill chain. When we look at what qualities created the desired effects in the decision cycle that Lt Col Shanahan described, it was the ability to orient fast and decide even faster. Staying in line with our kill chain model, we should maximize our F2T speed as well as the speed with which we transition to Target. Both measures of performance relate to speed but require different kinds of improvements to ABMS. Identifying avenues for these improvements does not require us to forge new fronts but rather to examine our successes and to rethink our current efforts. Air Force Chief Architect, Preston Dunlap, champions the value of the “10 percent or 20 percent solutions,” for which we can offer proofs-of-concept. From these little wins, we can start to see where we can focus our efforts to maximize the value already starting to arrive to the ABMS sensing grid. Lean Toward Resources in the Private Sector Being able to shift our weight from the F2T phases into the Target phase is the pivot point that would most likely disrupt the enemy kill chain as they see us recognizing their intent and posturing to counter it before they can act. The core competency to us being able to move to targeting is the ability to communicate our targeting intent to our operating assets in the sky. Some of our biggest wins in this arena have been enabled by our partnerships with the private sector. In December of 2020, the Air Force successfully tested a “translator” called gatewayONE. It was designed to enable communications between the F-22 and F-35 so the aircraft could share data with each other without giving up their position. This was made possible through the collaboration of Lockheed Martin, Northrop Grumman, and Honeywell who leveraged their collective expertise to fashion a proprietary antenna solution to make it all possible. Additionally, in 2019, the Air Force Research Laboratory worked with SpaceX to test Starlink’s capability to connect to airborne assets. While beyond line-of-sight capabilities have existed for a long time, SpaceX’s Starlink improved upon that capability by integrating both a CJ12 and an AC-130J gunship. This proved to provide a higher data access speed than that of its usual connection medium. These are examples of leaps forward that we made which may not yet improve our direct lethality on the battlefield, but certainly represent steps in the right direction. These capabilities bolster the methods and means with which we can pass targeting information to our operating ends, thus improving the speed with which we can pivot from our F2T phases to the Target phase. These wins have only been made possible by empowering the Air Force’s greatest asset, its Airmen, to link with the best partnerships, expertise, and resources that our nation’s industry has to offer and the Air Force should continue to seek more wins this way. Prep Data for AI, Not AI for Data It’s been no that the key to the next step in Intelligence, Surveillance, and Reconnaissance (ISR) dominance is Artificial Intelligence (AI) and Machine Learning (ML). Lt Col Shanahan states that “If Newtonian momentum is defined as mass multiplied by velocity, we best define agility in the information age as movement multiplied by intelligence.” AI/ML stands to be one of our biggest force multipliers in maximizing the speed and number of options we can provide to commanders before we make that key pivot into the Target phase. There are a plethora of measures the Department of the Air Force is taking to rise to the challenge of AI/ML but, as we do so, we need to make sure we are also prepping the data space for getting the maximum value from the AI/ML that the Air Force is funding. In his article The Way We Train AI is Fundamentally Flawed, Will Heaven describes a problem called “data shift.” Data shift is where AI/ML developed to pristine standards in the lab fails to follow through on its functions in real settings because of the “mismatch between the data the AI was trained and tested on and the data it encounters in the real world” (see Figure 2). These mismatches can often be attributed to incorrect assumptions and understandings that the developers had regarding the kind of data their models would encounter. Figure 2. Data Disparity in AI Development Herein lies a potential pitfall that we may soon find ourselves in. As we speed to build our AI/ML capes, we need to ask ourselves not if AI is ready for our data, but if our data is ready for AI. Are developers acquainted with all the different data standards, metadata, and data types that the Air Force operating environment produces? Are our databases built for enabling decision advantage or are they just specializing in the correlation of data? These are all, of course, cumbersome questions but they can be exposed and approached in a methodical manner. Our offering is that, given the vast amounts of data ABMS will operate on, ABMS itself needs a Data Strategy that inherits concepts from the USAF Data Strategy. This would give strategic level guidance that can then be interpreted at the operational and tactical levels to prepare the data space and handle new acquisitions to meet our prospective AI/ML efforts. Many different factors can play into speeding up our ability to manage F2T, but few pose a better chance to provide a new offset quite like AI/ML does. This simple mindset shift could set the USAF on rails to develop the speed we need in F2T and disrupt adversary kill chains. Conclusion Deriving value from the lessons in our past, we see that decision advantage can be provided in competition through maximizing the capability of our F2T phases of ABMS and being able to pivot into our Target phase in a manner that disrupts the enemy kill chain and potentially traps them in their own F2T cycle. Much of the good work going on for ABMS lately has been toward the end of “more connection” and “more speed.” One could not select truer avenues for improvement in warfare but, with the analysis provided, we may finally benchmark a goal for our speed and connectivity that has a fidelity higher than just “more.” By maximizing private sector partnerships and reorienting our understanding of how to effectively approach AI, we stand to achieve a good first goalpost for the ABMS sensing grid as we continue to build and evolve the process to maintain air dominance in both conflict and competition. Captain Alexander Wright Captain Wright is a Network Operations Officer currently assigned as a Cyber Operations Planner on the Secretary of the Air Force’s Staff, Studies and Analyses Directorate, Pentagon where he works closely with the Advanced Battle Management System (ABMS) leadership team to characterize system architecture and operations for ABMS’s ongoing development. His previous assignments include Chief of Cyber-Intel Integrations for the 693 Intelligence, Surveillance, Reconnaissance Group, Flight Commander of Mission Systems for 693 Intelligence Support Squadron, and Information System Security Manager for the Korean Air Operations Center. Captain Zachary Munoz Captain Munoz is a Defensive Cyberspace Operations officer assigned as Chief, Standardization and Evaluations at 834th Cyberspace Operations Squadron, Joint Base San Antonio-Lackland, Texas. Other previous assignments include Team Lead, 801 Cyber Protection Team, Joint Base San Antonio-Lackland Texas and as the Officer in Charge of Cybersecurity for Early Warning Radars, Air Force Life Cycle Management Center, Peterson Space Force Base, Colorado. Captain Seamus Feeley Captain Feeley is an Intelligence Officer and Mission Commander currently assigned to the 6th Special Operations Squadron where he conducts Unconventional Warfare and Foreign Internal Defense taskings, worldwide. He commissioned through Officer Training School in 2016 and has a bachelor's degree in Intelligence Studies and a master's in management. Captain Timothy M. Bjorgan Captain Bjorgan is a B-1 Weapon Systems Officer (WSO) and currently serves as the Chief of Weapons and Tactics Flight for the 34th Bomb Squadron at Ellsworth AFB, South Dakota. Previous assignments include Scenario Development Lead for the Weapons and Tactics Flight and Scheduling Assistant Flight Commander, both for the 34th Bomb Squadron. Captain Shelby A. Copenhaver Captain Copenhaver is an Instructor Weapon Systems Officer on the AC-130J Gunship, assigned to the 73d Special Operations Squadron, Hurlburt Field, FL. She has deployed twice (2019 and 2020) to Afghanistan in support of Operation FREEDOM'S SENTINEL and the RESOLUTE SUPPORT mission, and is a distinguished graduate of Squadron Officer School (2021). Captain Jacob Garrett Capt Garrett is currently assigned as an Instructor Air Battle Manager at Undergraduate Air Battle Manger Training at 337 Air Control Squadron, Tyndall AFB. Previous assignments include Deputy Chief Weapons and Tactics as a Senior Director and Evaluator Air Weapons Officer, 12 Airborne Command and Control Squadron, Robins AFB, GA. This paper was written as part of the SOS Air University Advanced Research (AUAR) elective on ABMS. NOTES [1.] Shanahan, John N.T. “Shock-Based Operations: New Wine in an Old Jar.” Air University, May 2, 2001, https://www.airuniversity.af.edu/Portals/10/ASPJ/journals/Chronicles/shanahan.pdf. [2.] Joint Lessons Learned Division, Figure 9. The Competition Continuum, image, DocPlayer, January 2019, https://docplayer.net/136833966-U-competing-in-the-information-environment.html. [3.] Hitchens, Theresa. “ABMS Demos Speed New Capabilities to Warfighters.” Breaking Defense, January 22, 2020, https://breakingdefense.com/2020/01/abms-demos-speed-new-capabilities-to-warfighters/. [4.] Heaven, Will Douglas. “The Way We Train Ai Is Fundamentally Flawed.” MIT Technology Review, November 18, 2020, https://www.technologyreview.com/2020/11/18/1012234/training-machine-learning-broken-real-world-heath-nlp-computer-vision/. [5.] NIX United, May 14, 2021, https://nix-united.com/blog/artificial-intelligence-vs-machine-learning-vs-deep-learning-explaining-the-difference/.