By Aaron Bazin
[Note: This article was originally published by The Strategy Bridge on March 21, 2017]
From television shows like Westworld to movies like Rogue One, practical and ethical issues surrounding artificial intelligence (AI) seem to be on the minds of many. From one perspective, the use of AI in war is just one more step in the evolution of technology, no different than the emergence of tanks, airplanes, nuclear weapons, or the Internet. According to experts such Bill Gates, Elon Musk, and Stephen Hawking, AI has the potential to be vastly different than any technology that has ever existed and could change every facet of contemporary society, including armed conflict. It is prudent to discuss how the military could use AI to improve its ability to fight and how AI could perhaps change the nature of war itself.
A QUESTION OF JUST AND DISCRIMINATE FORCE
In many ways, military forces using AI on the battlefield is not new at all. At a simplistic level, the landmine is perhaps a good starting example. The first known record of landmines was in the 13th Century in China and they emerged in Europe somewhere between 1500 and 1600. Most landmines are not intelligent and all and apply a binary logic of “kill” or “don’t kill.” What landmines lack, and one of the primary reasons they are banned by most countries, is the ability to use just and discriminate force. As far as computers have come since the British used “The Bombe” to break the Enigma code, the human mind still has an advantage in determining the just and discriminate use of force and thinking divergently about the second and third order effects resulting from the use of force. But, according to some, that advantage may not last for long.
Today, there is a debate raging as to whether computers will ever be as smart as humans. According to some experts, such as Kurzweil, Horn, and Moravec computers could have the processing power of the human brain somewhere around 2030. There are others, such as Allen and Nicolelis, who have their doubts and disagree with the idea that machines will ever become as “smart” as people. Despite what side of the debate you come down on, if the processing power of computers continues to grow as some have predicted, the potential exists for AI computing power to match or surpass the human brain.
The Evolution of Computing Power
When most people think about artificial intelligence at war they tend to think about how AI is represented in popular culture in movies such as 2001: A Space Odyssey, Terminator, or maybe Ex Machina. In these movies, independent computers become aware and then try to destroy their creators. Even though this plays to our greatest anthropomorphic fears, there is an emerging idea that the use of AI and human brain combined could make better decisions than either could alone. So instead of an either-or-proposition, as computers advance we should develop them to complement, not compete with the human. This is the main thrust of the idea behind the idea called intelligence augmentation. The military that first takes this approach and combines it with cognitive computing that already exists, could gain a decisive advantage on the battlefield.
USING “INTELLIGENCE AUGMENTATION” AND COGNITIVE COMPUTING TO ASSIST IN STRATEGIC DECISION-MAKING
For a moment, consider the relationship you most likely have with your smart phone. When you have trouble remembering a fact you need to know, it is there to augment your memory. When you are on vacation and are looking for the perfect restaurant to go to, it can help you decide which one to choose based on the mood that you are in. If you are looking to keep yourself entertained while you are waiting in the doctor’s office, your phone is there with a game. In many ways, the relationship you have with your smart phone is that it augments your own cognitive abilities. This relationship is very similar to the idea behind intelligence augmentation.
In the paper, Computing, Cognition and the Future of Knowing, John Kelly of IBM argues that the world is entering the era of cognitive computing where technology augments intelligence by improving “the human ability to understand – and act upon the complex systems of our society.” The relationship they describe is one where humans help build the corpus of knowledge for the cognitive computing system for a specific knowledge domain, interact with it to in effect train the system, and then work with the system to make informed, evidence based decisions.
Most of today’s computers follow programs based on the computation of mathematical solutions. When dealing with complex and novel situations, decision tree models of this sort breakdown. This is where cognitive computing comes in. Cognitive computing systems work in way similar to the human brain using pattern recognition and natural language processing (see video below). A cognitive computing system provides an advantage in that it can tap into stores of “big data” in a usable way for the human decision maker. This could enable a decision maker to tap into an ocean of data that would otherwise be hidden from them.
So far, the use of intelligence augmentation to address complex problems has shown promise. For example, augmented approaches to diagnosing cancer have proven much more accurate than relying solely on a doctor’s judgment. If the military chooses to take an “intelligence augmentation” approach and leverage cognitive computing, it may just gain a decisive advantage on the battlefield of the future. So the question is, how could the military leverage this emerging cognitive computing technology to enhance the ability of decision makers to address the challenges of an ambiguous, complex, and rapidly changing security environment?
AUGMENTED STRATEGIC ANALYSIS: A VIRTUAL CLAUSEWITZ?
If applied to its full potential, this technology could open up new opportunities for strategic leaders seeking to make sound and timely decisions on the battlefield and offer best military advice to political decision makers. Done right, this system could leverage a broad corpus of experience, education, and professional knowledge to augment a strategic leader’s ability to understand, identify options, and make evidence-based decisions. The experts could help build a corpus using primary sources and data from three broad areas: experience, education, and professional knowledge.
The system would then have to be trained through regular interaction with a pool of subject matter experts who help guide its development and assign more weight to the primary sources that matter most (e.g., Clausewitz, Sun Tzu, Boyd, etc.). Finally, the cognitive computing system would continue to learn and improve using as more leaders use it over time through both actively and passively. The diagram below depicts how the military could develop such a system to augment strategic analysis:
An Augmented Strategic Analysis Framework
In the experience category, the corpus could include data from every combat lesson learned ever captured, every after-action review ever conducted at every combat training center, and a detailed history of every battle ever recorded. It could also include the detailed memoirs of generals and politicians throughout history. The idea is that over time the system would learn patterns and grow the ability to determine what really matters. Of course, it would be difficult to convey the human feelings of frostbite in a foxhole as or what it means to lose a soldier as a battalion commander. The point is not to replace the human, but to arm the human decision maker with the best evidence upon which to base his or her decision.
In the education category of the corpus, the cognitive computing system would need both a full range of both civilian and professional military education. In the civilian category, the experts could help build a database on a broad range of topics that would comprise both a liberal arts and STEM undergraduate background. Then upon this, the next type of information would be graduate-level primary including international relations, political science, economics, psychology, philosophy, anthropology, and others. Finally, to round out the education category, every lesson from pre-commissioning to war college could be incorporated, including specialized military courses such as the Basic Strategic Arts Program and the School of Advanced Military Studies.
The final category is professional knowledge. This could include primary sources from professional reading lists, defense-related journals and websites, and academic papers. If desired, those wanting to solve a specific problem could upload vast quantity of data including detailed budget information, personnel records, or other technical information. Overall, this could never replace individual study, reflection, or research. The point is to augment, not replace the human. As events unfold across the world, the system could filter news reports and continually add them into the corpus to help the strategic thinker maintain situational awareness and understanding. As new ideas emerge in the profession and are discussed in professional forums they would be included in the corpus.
The corpus just forms the basis for the cognitive computing system; it would have to interact with a group of subject matter experts and then end-users. Subject matter experts must help curate the data and help the system improve to reach its full potential. Once complete, the system would act as the personal advisor answering requests for information and providing options. The system would base all the on a deep access to information of the phenomena of war as it has unfolded across history and make evidence-based assumptions about the future. Link this to an app on a smart phone or an Amazon Echo, and you have a virtual Clausewitz at your beck-and-call. Ultimately, the goal is that this system and the strategic leader would work together and be more effective than either could be working alone.
POTENTIAL BENEFITS AND LIMITATIONS
The Secretary of Defense, retired General James Mattis has said, “[the] most important six inches on the battlefield is between your ears.” If you believe this yourself, then you can see how cognitive computing has a tremendous potential help improve a military’s ability to fight and win wars. Strategic leaders are in the business of solving complex problems, making difficult decisions, mitigating risk, and providing unfettered advice. They need the ability to see the big picture as well as focus in on the details. A properly built and trained cognitive computing system of this type could help the military derive answers to strategic or operational questions grounded firmly in data. Arguably, the biggest benefit is that it could help a strategic leader to make decisions faster and better than their adversary, giving them the cognitive advantage on the battlefield.
During strategic-level decision making, cognitive biases can creep into thinking and lead to a distorted view of reality leading to miscalculation. Cognitive computing can be effective in helping a strategic leader identify the blind spots in their thinking and overcome biases that they, their staffs, and institutions hold near and dear. Militaries could use cognitive computing systems to wargame multiple courses of action ahead of time that could in turn identify novel and emerging solutions.
Scholars have argued that war and military conflict have an enduring nature with an ever-changing character. One example of the enduring nature is that war is fundamentally a human endeavor. Using a cognitive computing system to help fight wars has potential to change this. A cognitive computing system will never know what it feels like to be human and what it is like to build a career of combat experience. But the system does not have to. That is the benefit of having the human retain meaningful control. Ultimately, the human would still make the decisions that the system could not, including making the important ethical and moral choices.
The thought of a computer deeply involved in life or death decisions scares many people. However, these types of feelings are nothing new. During the Industrial Revolution, there were groups called Luddites who destroyed machines and used violence to oppose to technological change. This teaches us that dramatic change is often met with resistance, because of the uncertainty that often accompanies change. In the case of cognitive computing, there is the added threat to people’s perceptions of their own uniqueness and expert status. If the military took an intelligence augmentation approach, it would take time and leader emphasis to overcome the natural fear of involving a computer deeply in life-and-death decision-making.
CONCLUSION
Of course, Clausewitz is more than just a CD-ROM. It may be overly optimistic to imagine the creation of a system consisting of everything from military doctrine to philosophy that can help create novel solutions to complex military problems. Combat has a way of making things difficult, and factors such as limited time, lack of connectivity, and over-reliance on such a system could cause an epic failure on the battlefield. But here, the idea is not to replace the human decision maker out-right. The idea is to empower strategic leaders with a tool that not only contains deep professional knowledge but also helps them make evidence-based decisions given the fog, friction, and chance that exists in war. The rare military geniuses that emerge do so out of experience, aptitude, and a lot of hard work. Simply put, if done right, a tool of this type could help them do what they do better, perhaps enhancing the ability of commanders use their intuition, an enhanced coup d’oeil.
Whether computers could soon eclipse humans in their ability to process information or not, cognitive computing and intelligence augmentation could be leveraged today on the battlefield today. This article describes one way how the military could build a cognitive computing system that uses broad corpus of experience, education, and professional knowledge to augment a strategic leader’s ability to understand, identify options, and make evidence-based decisions. Intelligence augmentation is only one of many possible ways to approach the maturation of cognitive computing to get the most out of it on the battlefield of the future. Just how this technology will change war and armed conflict remains to be seen.
Comments