Intervention and the Looming Choices of Autonomous Warfighting
On July 1, the Office of the Director of National Intelligence revealed that the United States conducted 473 “U.S. Counterterrorism Strikes Outside Areas of Active Hostilities” between January 20, 2009 and December 31, 2015. Here’s something that wasn’t in the release — record of a single American pilot’s death. The use of drone aircraft permits policymakers to exercise the use of force abroad while substantially lowering the risk to military personnel, potentially lowering the threshold for use of force.
Recently on War on the Rocks, various thinkers from military circles exchanged views on what ethical and legal limitations should constrain autonomous warfighting systems and interpretations of DoD Directive 3000.09. But another question looms. Impending advances in artificial intelligence and renewable energy technology will result in semi-autonomous weapons systems capable of performing an increasing number of missions without human support, beyond precision strike. If remotely piloted drones lower thresholds for use of force now, will growth in machine independence lower thresholds for foreign intervention?
Greater Self-Reliance, Range, and Deniability
Human combatants require food, water, sleep, and fuel to fight, but robots require only fuel. As commercial industry increases the collection efficiency of renewable energy platforms and capacity of batteries, the combat endurance of military robotic systems will continue to grow. The MQ-9 Reaper can fly 1,000 nautical miles before refueling, but future drone aircraft may be able to use solar power to stay aloft until they require human maintenance. Sea-going robots will have the option of at least two energy systems: solar power and the motion of the ocean itself. Adaptable drone systems could even switch between these sources based on environmental factors. And like the iRobot Roomba that vacuums my house, these semi-autonomous systems will possess the intelligence to periodically charge their batteries through periods of mid-mission rest. Assigning teams of robots to the same task will ensure continuous coverage, even as other robots recharge themselves.
While advances in the energy sciences will keep robots working longer, simultaneous advances in object recognition and miniaturized computer processing power will allow robotic systems to maneuver and act with fewer human inputs. Today’s precision-guided munitions (PGMs) and drones rely primarily on satellite-based navigation systems and inertial navigation backups, but future integration of internal lasers, radars, and cameras will allow both drones and PGMs to maneuver autonomously, just like Google’s self-driving car. When these weapons arrive at their destination, they will locate targets though object recognition by comparing the features of objects in their environment to internal databases of enemy systems. Algorithms may even grow precise enough to recognize enemy camouflage patterns or individual weapon systems. As a safeguard against civilian casualties and fratricide, commanders could draw geographic bounding boxes within which drone systems classify all parties as friendly or hostile — similar to the permissive and restrictive fire support control measures the U.S. military uses today.
These changes will allow nations to deploy semi-autonomous systems farther away than current technology allows — the MQ-9 Reaper’s range necessitates that logistical bases stay relatively close to targets — and increase the ability of nations to claim plausible deniability if they employ drones in covert actions.
These developments in robotic endurance, target acquisition, and navigational autonomy will manifest in systems on the air, ground, and sea that provide robust capabilities to accomplish missions beyond intelligence collection and precision strike.
Counter-Piracy: At key locations, such as the Horn of Africa, nations could deploy swarms of autonomous patrol boats programmed to accompany civilian ships through dangerous areas. A single naval vessel could exercise command of all drones in the area and make use-of-force decisions based on the monitoring of video feeds and radio conversations with civilian crews concerning the threat posed by potential pirates. Solar- or wave-powered charging stations for drones could be built in the area.
Blockade: Energy self-sufficient “smart mines” could be emplaced around the harbors or key water ways of adversaries. With organic target acquisition systems (radar, sonar, etc) these systems could autonomously detect ship movement, use algorithms to identify enemy ship types, and launch torpedoes or anti-ship missiles with their own object-recognition guidance systems. Dormant mines could lie around potential zones of conflict for months at a time, trickle-charging their batteries through a solar link to the surface or by recycling the energy of currents. Networked smart mines could create redundant, aggressive systems that would pose enormous challenges for traditional minesweepers.
Base Defense: A collection of semi-autonomous systems could reduce the number of human beings required for base defense. Sentry guns, capable of identifying adult males, could line forward operating base (FOB) walls to deter ground assault, and airborne drones could conduct the equivalent of counter-mortar patrols. Air defense systems would detect incoming projectiles and autonomously deploy interceptors to mitigate threats. These network defenses could be monitored from a FOB security center where a base commander would make decisions concerning use of force and deployment of human troops.
Persistent Intelligence Collection: More energy-efficient intelligence collection systems and reconnaissance assets will be able to remain in target areas for days, weeks, or months at a time. Their endurance will reduce gaps in collection coverage and could allow terrestrial assets to supplement or replace collection currently undertaken by space-based entities.
Used in conjunction, these capabilities could significantly reduce the number of military personnel or manned assets needed to perform complex military missions, which would potentially lower domestic political risks associated with expeditionary military missions.
Comparative Military Advantage
But will these technological innovations result in lower thresholds for the use of force and intervention abroad? It depends on the opponent.
The United States has used its air and naval superiority to conduct punitive strikes within Afghanistan, Libya, Iraq, Serbia, and Sudan without foreign permission, but never against near-peers like Russia or China. Semi-autonomous systems provide a technological avenue for the United States to expand the variety of missions it conducts abroad while exposing fewer service members to danger, but they will likely not be used against nations that possess credible means of retaliation. Future use of semi-autonomous systems poses the biggest risk to the same nations threatened by Tomahawk cruise missiles today, those too weak to project power back.
We can expect policymakers to exploit the capability gaps created by semi-autonomous systems, but the advantage will be fleeting, at least against sophisticated adversaries. After the successful demonstration of a technology, the clock begins. Other nations will rush to develop or acquire comparable abilities, align with powers who can, or begin developing countermeasures to defeat or limit the technology’s effectiveness. The success of the U.S. drone program has already spurred over 80 countries to develop their own. Future successes will undoubtedly inspire more emulation.
Whatever your feelings about the risks of war’s automation, countries who fail to adequately develop autonomous warfighting systems are most likely to be the victims of them. The autonomous arms race has begun. Unless a counter-proliferation framework is created, liberal societies can either participate or risk the loss of a military advantage that might one day endanger their sovereignty.
Jules Hurst is a Georgetown graduate student, army reserve officer and the former senior intelligence analyst of 1st Ranger Battalion. He has deployed to Afghanistan four times in support of a USSOCOM Task Force. You can follow him @JulesHawke. The views expressed here are his own and do not reflect the views of the Department of Defense or the Department of the Army.