Dr. Robert J. Bunker
Key Insights:
- A number of definitions for autonomous systems exist. One of the working definitions used at the colloquium was that in which their key feature is viewed as the ability of a given system to perform certain functions independently. Autonomous systems can support military operations by contributing to force protection as well as increasing mobility and improving lethality. The level of autonomy varies across platforms, but human oversight is still viewed as central to the deployment of autonomous systems in military operations.
- The definition of the term “robot,” on the other hand, is generally more agreed upon. It signifies a technology that requires sensors, artificial intelligence, and tools to carry out its tasks. However, much debate still exists concerning human decisionmaking requirements—vis-à-vis in the loop, on the loop, or outside the loop—with regard to armed robotic systems. This debate is framed by increasing the observe, orient, decide and act (OODA) loop requirements in warfare that surpass human cognition capacity (military necessity) weighed against the traditional view that only human beings should be responsible for the decision to take the lives of other human beings (ethics and morality).
- Limitations on autonomous robots are derived from mechanical (effectors), environment (hazards), and mission (objectives) variables. Simple robot use scenarios are far more favorable to autonomous systems because the complexity threshold is lower. Therefore, a sense-model-plan-act (SMPA) model is utilized to contend with the robotics complexity problem. This model is based on SMPA iterations and draws upon deliberate and reactive acts. This is why open air and open sea type environments are much easier at present to operate autonomous systems in than complex and populated urban environments. Emerging technologies that may overcome these limitations are based on probabilistic, networking, and parallel processing innovations.
- Our adversaries—nonstate and even potentially state based—will not be constrained by our democratic legal and ethical inhibitors in using these systems. Many of these will be enemies of all civilized peoples. For some years now, groups such as Hezbollah, al-Qaeda, Islamic State, and others have been utilizing unmanned aerial vehicles (UAVs) for terrorist plots, anti-personnel targeting, reconnaissance, and even propaganda purposes. As a result, we need to not only red team current and near future opposing force capabilities (to about 10 years out) but to develop countermeasures to opposing remote controlled and semi- and fully-autonomous systems (e.g., countercyber control targeting) and operating concepts (e.g. counterswarm).
- Continuity and change issues must also be considered when looking at these systems. Robotics and autonomous systems (RAS) will not be revolutionary in the sense that they will change the fundamental nature of war. Four constants in war exist; war is an extension of politics, it is profoundly human (fear, honor, and interest), it is characterized by uncertainty (complexity, interaction with enemies, and new technologies), and it is inherently a contest of wills. Still, change from an evolutionary perspective will be profound, similar to that of the revolution in military affairs (RMA). From the context of land warfare—that is, from a human muddy boots on the ground perspective—this change will be less significant, however, than it will be in naval, air, and space mediums.
- Robotics and autonomous systems technologies quite likely exist within the context of much greater revolutionary potentials intertwined with nonlethal targeting, new energy sources, advanced manufacturing (3-D and 4-D printing), networked and cloud based information and social media, the commercialization of security, and changing human values. U.S. and Canadian militaries are compelled to innovate RAS before being forced on the battlefield to do so. Potential peer competitors and a rising power, such as China, now have plans to produce fleets of armed drones that could operate in swarms of thousands, and even tens of thousands, of units. Conventional ground fighting vehicles, aircraft, and warships would be asymmetrically challenged by far more advanced military technologies such as RAS swarms.
- Land warfighting challenges and the impacts of autonomous systems will likely be clustered in the areas of a) reconnaissance capabilities, b) securing wide areas across the battlefield, c) sustaining Army and joint forces across that battlefield, and d) influencing joint combined arms and multinational operations with new swarming and other evolving maneuver and fires capabilities.
The 10th annual Kingston Conference on International Security (KCIS) was held in Kingston, Canada, May 11-13, 2015. This significant North American military research activity represents an important bilateral academic strategic outreach interchange between Canadian and American Landpower forces and has been held since 2006. This year’s conference theme focused on “Robotics and Military Operations” and was co-sponsored by the Centre for International and Defence Policy, Queen’s University; the Canadian Army Doctrine and Training Centre of the Canadian Forces, Royal Military College of Canada; the Canadian Army Command and Staff College; and the North Atlantic Treaty Organization (NATO) Defense College; and was conducted in cooperation with the Strategic Studies Institute of the U.S. Army War College. The 162 conference participants and attendees were drawn from government, nongovernment organizations (NGOs), industry, academic, and military entities from across the United States, Canada, and NATO.
Robotics and military operations, as a colloquial focus, span remote controlled, semi-autonomous, and autonomous systems utilized by both allied and opposing forces along the entire conflict spectrum and across battlefield domains (e.g., land, air, sea, space, and cyber). The conference was divided into seven panels:
- State of the Art: Current and Emerging Technologies;
- State of Play: RAS in Recent Operations (Allies and Adversaries);
- State of Governance: Law and Policy;
- Ethical Implications;
- Assessing, Detecting and Responding to RAS Threats;
- Force Development Strategies: Revolution vs Evolution; and,
- Policy Recommendations.
Dr. Peter W. Singer, of the New American Foundation and author of the influential work, Wired for War (2009), delivered the keynote address titled, “Robots, Autonomy, and the Next World War.”
Some of the key military documents related to the environments in which robotics and autonomous systems will be utilized are; TRADOC Pamphlet 525-3-1, The U.S. Army Operating Concept: Win in a Complex World 2020-2040, published in the United States, and The Future Security Environment 2008-2030: Part 1: Current and Emerging Trends (Chief of Force Development), published in Canada. Specific military documents concerning RAS fielding and evolving doctrinal thinking include Department of Defense Directive 3000.09, Autonomy in Weapon Systems (U.S. Department of Defense); Fiscal Year 2009-2034 Unmanned Systems Integrated Roadmap (U.S. Department of Defense); Chapter 2, “Robots in the Army,” and Chapter 3, “Army Robotics Concept,” No Man’s Land: Tech Considerations for Canada’s Future Army (Canadian Army Land Warfare Centre); and the expected publication of the U.S. Army, Remote and Autonomous Systems Strategy, in early 2016.
No legal prohibition of any semi or actual autonomous weapons systems in international law exists. The proposed preemptive “killer robot” ban by some NGOs (e.g. the 2012 Losing Humanity report and ensuing lobbying effort) has no legal standing. This interpretation is derived from a traditional Laws of Armed Conflict (LOAC) review based on effects and results as determined by the answers to the following questions: Is it a lawful target? Is the weapon carried legal? What is the probability of collateral damage? If the answers to the first two questions are yes and the attack does not produce excessive collateral damage, then an attack by a semi or actual autonomous weapon is legitimate. Even so, efforts must be made to minimize collateral damage. Given these caveats, attacks by semi or actual autonomous weapons are perceived as moral and legal and can be employed, but constant care must be maintained as targeting/use circumstances change. Therefore, no “meaningful human control” governor is required by LOAC requirements on such autonomous systems.
Shrinking military manpower and the desire to minimize battlefield casualties are expanding the fielding of "bots on the ground" as "boots on the ground" decrease. Along with the expected cost savings of fielding combined human and robot armies by Western militaries, RAS technologies allow for space to be created between shouting and shooting responses. A U.S. Air Force UAV commander cautions against underestimating manpower costs associated with employing robotic systems. Actual manpower costs of employing UAVs around the world exceed the number required for employing manned platforms. By combining robots with nonlethal weaponry, more precisely graduated responses can be achieved in military operations other than war, such as in peace maintenance missions. Additionally, deterrence may become a more effective factor due to the improved indications and warnings (I&W) that these systems can provide. However, the possibility of rapid escalation must be addressed due to precisely graded responses quickly escalating a conflict via many incremental actions taking place too fast due to the compressed OODA loop cycles of opposing forces.
A number of the presentations and discussions that took place during the colloquium resulted in various emerging strange new world scenarios. In the increasingly high-tech world of the latter-21st century, the avatar scenario (of remotely controlled systems) called Track I was contrasted with the terminator and/or robocop scenario of fully autonomous malevolent and/or benevolent artificial intelligence (AI) systems called Track II. Between these extremes existed the cyborg—human and machine integration and singularity—scenarios along with “an internet of things” network scenario, existing simultaneously in cyberspace and in massive robotic swarms of small expendable devices. This resulted in some far-ranging debates, quite a few of which focused on Track II technology. Debate centered on whether true AI (machine awareness of self) could ever be achieved and, if so, would humans really want to arm such sentient machines and robotic AI swarms (a human security approach).
Currently, the Canadian Army is addressing the following four major questions that pertain to RAS threats: 1) How will RAS adversary systems—either nonlethal or lethal—affect our own tactical operations? 2) How should these threats be anticipated given technical, legal, moral, and ethical constraints of state, nonstate groups, and individual actors? 3) How will unstructured/complex environments and unpredictable/evolving scenarios affect our ability to respond? and, 4) How should Canada approach the development of capabilities to counter such systems?
Distinguished speakers included: Lieutenant General Guy Thibault, Vice Chief of Defence Staff, Canadian Army; Lieutenant General H. R. McMaster, Director, U.S. Army Capabilities Integration Center; Major General Jean-Marc Lanthier, Commander, Canadian Army Doctrine and Training Centre; Major General Stephen Bowes, Chief of Force Development, Canadian Armed Forces; Major General Robert Dyess, Chief of Force Development, U.S. Army; Professor Stéfanie von Hlatky, Centre for International and Defence Policy, Queen’s University; Dr. Jeffrey Larsen, NATO Defense College; Professor Douglas Lovelace, Director, Strategic Studies Institute, U.S. Army War College; and Professor Steve Metz, Director of Research, Strategic Studies Institute, U.S. Army War College.
The conference was a superb example of scholar-practitioner and academic-military interactions and helped to promote the development and sharing of new concepts concerning RAS development, fielding, policy, operations, and countermeasures. It also provided a venue for international exchange between the United States, Canada, and NATO in a cost-effective manner. Further, partially derived from ongoing U.S. and Canadian bilateral requirements identified during last year’s conference, senior U.S. defense leadership fully re-engaged this year with this important conference series.
For additional information on the Kingston Conference on International Security and next year’s annual event, contact the conference coordinator, Ms. Maureen Bartram. For more encompassing conference information, select presentations will be posted at the conference website, queensu.ca/kcis/index.html. An archival Twitter feed also exists via QueensCIDP #KCIS2015.
*****
The views expressed in this Strategic Insights article are those of the author and do not necessarily reflect the official policy or position of the Department of the Army, the Department of Defense, or the U.S. Government. This article is cleared for public release; distribution is unlimited.
*****
Organizations interested in reprinting this or other SSI and USAWC Press articles should contact the Editor for Production via e-mail at SSI_Publishing@conus.army.mil. All organizations granted this right must include the following statement: “Reprinted with permission of the Strategic Studies Institute and U.S. Army War College Press, U.S. Army War College.”