Is the dawn of artificial intelligence and autonomous weapons resulting in a new revolution in military affairs? That’s the question posed by a recent article by Frank Hoffman. As with previous such revolutions, technology is not the sole determinant of large shifts in the way military organizations successfully conduct operations. Development of new organizations and new warfighting concepts also play a large role — and these are the domain of humans, not machines. Therefore, if military organizations are to successfully adapt to this new era, they must maximize their human potential to gain an intellectual edge. To sustain and build this edge, leaders will need to rethink professional military education systems and learn how to best combine human and artificial intelligence.
This article will explore how military organizations can ensure that its people are intellectually prepared for success in future warfare by using the potential of artificial intelligence, as well as how this may impact professional military education. If done early enough, these changes to the educational model will allow institutions to maximize the strengths of humans and artificial intelligence in a collaborative environment across the military enterprise. In doing this, organizations may also unearth new and more effective ways of thinking about, planning, and executing military operations.
The Military Potential of Artificial Intelligence
Military institutions, policymakers, and intelligence agencies are feeling the competitive pressure to expand the use of military (and broader national security) applications of artificial intelligence. Both the U.S. and Chinese governments have released long-term strategies to lead the world in the development and employment of artificial intelligence.
But what is artificial intelligence, or AI? As Michael Horowitz noted in a recent article in the Texas National Security Review, there is no universal agreement on the specific meaning of the term. For our purposes, I use the definition posited by Amir Husain in The Sentient Machine: Artificial intelligence is the overarching science that is concerned with algorithms, whether they learn or not from data. Building on this definition, Genevieve Bell of the Australian National University describes AI as “just the 21st century steam engine. It is not the end game. But it is the means to getting things done.” Finally, the potential applications described in this article assume the use of the “narrow” artificial intelligence that is currently within the technological capabilities of Western nations.
A principal reason that militaries need artificial intelligence is the convergence of large quantities of sensors, communications networks, and an accelerating stream of data and information. As the quantity of information continues to increase, the capacity of humans to deal with it is not increasing commensurately. Indeed, humans are fast becoming the most sluggish link in decision-making. And while there is much ethical debate in the West about the application of autonomous weapon systems, as Ian Morris has written, “when robots with OODA [observe, orient, decide, and act] loops of nanoseconds start killing humans with OODA loops of milliseconds, there will be no more debate.”
Insights into the applications of artificial intelligence are widely available from civilian businesses. Microsoft’s Cortana, Google’s Now, Amazon’s shopping sites, Netflix’s streaming algorithms and Apple’s Siri all combine user input, access to large databases, and bespoke algorithms to provide decision support for human users based on their previous decisions and those of millions of other users. This represents, as Yuval Harari has described, a slow and steady shift of authority from humans to algorithms. There are pitfalls in moving some decision support and decision-making to machines. But this move also offers military organizations potential benefits in information filtering, exploring more options, and deeper analysis of those options. This may provide for faster, or better-quality, decision-making in complex operational and strategic environments.
Artificial Intelligence and Intellectual Preparation for War
To begin intellectually equipping their personnel for the Fourth Industrial Revolution, military organizations should start providing a baseline level of education in artificial intelligence. This intersection of artificial intelligence and professional military education would foster broader literacy about the concept among military personnel, integrate artificial intelligence into the teaching process, and as a recent article by Benjamin Jensen, Scott Cuomo, and Chris Whyte explored, use it to assist military institutions in exploring new warfighting concepts.
Teaching Artificial Intelligence
Just as AI has proliferated in commerce, so too is it likely to proliferate through military decision-making. It may be used in intelligence analysis, strategic decision support, operational planning, command and control, logistics, and weapon systems across all environments. But if institutions are to use AI, they will need informed users to shape the application of AI and provide quality assurance. A challenge with AI is its lack of “explainability.” It is not always obvious what the algorithms that underpin these new capabilities do. This “black box” quality should drive a different approach to preparing military personnel for its various applications.
Military organizations will therefore need more than just deep technical experts in the development of algorithms and the design of artificial intelligence for military systems. As a recent U.K. government reporthas found, any skilled workforce using AI should be a mix of those with a basic understanding, more informed users, and specialists with advanced skills. Over the coming years, at almost every rank level, military personnel will require basic literacy in artificial intelligence, including knowledge of its application, how to provide a level of assurance and quality control, and how to optimally combine it with human intelligence. Western military education systems currently do not provide this enhanced literacy for all their personnel. However, as the U.K. report notes, it is the coupling of technical experts with a heightened technological literacy across the entire force that will allow future military organizations to fully exploit the benefits of artificial intelligence.
As militaries seek to develop a baseline literacy in AI to supplement the knowledge of experts and contractors who design algorithms, reading lists, residential programs, and online skills as well as academic partnerships and conferences will be useful. Existing professional military education programs must be adapted to deliver this higher standard of technical literacy. Currently, at least in the West, professional military education curricula generally focus on three key areas: national security policy and strategy, joint warfighting, and the important area of command, leadership, and ethics. While there are specialist courses offered in procurement and logistics, the opportunities to develop broader understandings of the implications and exploitation of advanced technologies are limited. Adding another pillar to curricula focused on technological literacy, the ethics of advanced technology, and procurement and logistics will create a wider institutional capacity to understand new technologies, foster quality control, and address the risks of bias and misbehaving algorithms.
This additional educational focus, continuous from the earliest days of an individual’s military service through to senior ranks, can help militaries develop a deep institutional reservoir of people who understand artificial intelligence and appreciate humans can collaborate with machines effectively at each level of command. But such changes are not simple. As with all significant institutional change, evolving professional military education to incorporate this additional area of intellectual focus will demand senior leader advocacy, investment in new curriculum, and disciplined implementation across joint and single service educational programs.
Learning with Artificial Intelligence
Artificial intelligence also has significant potential to change the way militaries educate their personnel. As one recent report described, artificial intelligence-driven intelligent tutoring systems may be able to provide simulated one-on-one human tutoring. Students might also benefit from an artificial intelligence “lifelong learning partner” that they are issued on entry into the military and that accompanies them through their career, a possibility that lines up nicely with the proposals Jensen, Cuomo, and Whyte make in their War on the Rocks article. For their part, instructors at military institutions may also benefit from their own teaching assistant that would communicate with students’ partners to interpret individual students’ profiles and provide suggestions on tailored learning.
Artificial intelligence also has application in collaborative learning, which military organizations have long used to share historical lessons and build effective and cohesive teams. Today, most collaborative learning is used for training (such as lower-level tactical skillsets). However, AI may also be able to help develop the cognitive skills that underpin higher-level operational and strategic planning in teams. It can do so by offering more authentic environments for collaborative learning, providing more intelligent adversary systems to challenge students, or using purpose-designed algorithms and curriculum data to amalgamate lessons from previous activities.
Finally, as Jensen, Cuomo, and Whyte described, artificial intelligence can be used for advanced simulations in professional military education. Currently, most simulation systems in military institutions are focused on training outcomes, such as gunnery, driving, flying, ships bridges, and artillery observation. There are some simulations focused on higher-level cognitive skills, such as the Joint Seminar Wargaming Adjudication Tool, the Joint Conflict and Tactical Simulation, and the Warfighters Simulation (WARSIM 2000). These tools, which generally require large numbers of interactors to provide background input, represent early efforts to educate more senior officers in joint and coalition warfighting. More sophisticated tools using AI could hone the cognitive skills of senior military leaders and, potentially, civilian policymakers.
Eventually, by linking human resource databases, doctrinal libraries, lessons learned databases, and curriculum designs, artificial intelligence may be able to identify gaps in the learning experience of military professionals as well as solutions to ensure they remain intellectually fit for future conflict. This may ultimately result in fewer bureaucratic distinctions between service professional military education and joint professional military education, enabling a single, unified education continuum across military and national security institutions.
Education and Artificial Intelligence in Concept Development
If military institutions can harness the benefits of artificial intelligence to build literacy about the concept and assist learners more generally, this can build the foundation for experimentation in new warfighting concepts for the 21st century. As Ellie Bartels has noted, wargaming is enjoying a renaissance in the U.S. Department of Defense. Multiple military educational institutions, such as the National Defense University and Naval War College, have deep expertise in helping their armed forces develop new operational approaches. There is also a long history of military educational institutions being used to develop future warfighting concepts — a key element of any transformation in warfighting capability.
Well-educated and experienced military officers conversant in the strengths and weaknesses of artificial intelligence will allow institutions to construct new, joint operating concepts that feature human-machine and human-artificial intelligence teaming. Through these explorations, organizations may find that the speed at which autonomous systems function forces humans to operate further up the chain of command. As Thomas Adams noted in 2001, tactical warfare may become the business of machines and not appropriate for people at all. And even as these experiments inform institutional modernization and operational concepts, they can also provide feedback on more effective and ethical use of artificial intelligence in the future.
These experiments might also suggest potential reforms to equipment, force structure, and personnel and career management structures. As Bell has noted, past industrial revolutions created jobs that didn’t exist before. It is likely that robust experimentation with the application of AI will render many existing military employment categories less useful and highlight the need for new categories. In 1913, there was no such category as “tank crewman,” but there were many horse-mounted cavalrymen. In 1945, we did not imagine “cyber warriors” as a core military capability. There will be future personnel categories we probably have not yet imagined that we will need to thrive in the digital age.
The U.S. Department of Defense, in a 2017 study of the future operating environment, described artificial intelligence as “the most disruptive technology of our time.” The potential applications of artificial intelligence and deep learning capabilities may be one of the most profound aspects of the Fourth Industrial Revolution. Indeed, for the first time in two centuries, informed and experienced experts and practitioners are questioning whether a technological shift will result in changes in the nature of war.
In future conflicts, decision cycles are likely to become faster than the capacity of human cognition to process. Military command and control and strategic decision-makers alike will need artificial intelligence that can process information and recommend options for making decisions faster (or of higher quality) than an adversary can. As I have written previously, military organizations will likely contain thousands or even tens of thousands of unmanned and robotic systems, all-encompassing some form of artificial intelligence. In this environment, where all sides may possess artificial intelligence and autonomous systems, the race will go to the intellectually swift.
This new age of autonomous systems, and the resulting new organizations and warfighting concepts, will demand that a larger percentage of military personnel are literate in artificial intelligence and other advanced technologies. It will require new ways of educating people in the military. And the speed at which technology will evolve will also require frequent re-skilling and re-educating of personnel in an evolved professional military education system. In leveraging the potential capabilities of AI in an evolved approach to professional military education, we can build an intellectual edge over future adversaries.