Free beta access. 50 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

Setting ethical standards for use of military AI undermines US competitiveness and causes a switch to worse weapons

Yoo 17

(John, Emanuel S. Heller Professor of Law, University of California, Berkeley, School of Law; Visiting Scholar, American Enterprise Institute, “Embracing the Machines: Rationalist War and New Weapons Technologies”, California Law Review, Vol. 105, Issue 2) DB

THE DANGERS OF OVERREGULATING NEW WEAPONS TECHNOLOGIES These new types of weapons allow nations to coerce and pressure each other in novel ways. This Part argues that banning these new weapons will prove futile. It will first base its analysis on a realist approach to international relations, which assumes that nations pursue their interest in security above all other interests and that cooperation will prove difficult due to the anarchic nature of the international system. It then shows why nations will not agree on the regulation of new weapons systems. It follows these predictions about international agreements by arguing that these new weapons may actually reduce the harm to both combatants and civilians in armed conflict. Counterintuitively, heavy regulation of drones or cyberweapons may have the unforeseen consequence of making war more dangerous and destructive. A. Realism and Weapons Before we can develop a sensible approach to regulation, we first must present a theory about international conflict. Without a theory, the international system might adopt standards without any understanding whether it is making the problems of war better or worse. Excessive regulation of new weapons technologies might unintentionally exacerbate the harms of war by forcing combatants to use more destructive weapons. Unduly lax standards might have the unintended consequence of allowing both sides to inflict grievous harms on each other to no military advantage. Without a theory of war, we do not know what values our legal rules should maximize and what costs it should minimize. We are not applying a philosopher’s attitude to war. Some believe that nations have followed laws during armed conflict out of a sense of morality that can trace its origins to Christian and, before that, Roman just war theories.81 We do not discount the intellectual tradition of just war theory or its recent appearances in the works of Michael Walzer, John Rawls, and others.82 But the history of war does not reveal a universal morality that imposes consistent restraints on warfare.83 This should come as no surprise in light of the diversity of the world’s religions and political, economic, and moral belief systems. Nations have often rejected moral appeals when they can achieve military advantage in combat. Instead, we follow a more instrumental approach. Like realist scholars in the field of international relations, we assume that the international system fundamentally suffers from anarchy. Anarchy does not mean that anything goes. Anarchy does not mean that the world lacks order, but only that no world authority can enforce international rules in the same way that domestic governments maintain law and order. The world is composed of independent sovereign states with no higher sovereign above them. Under realism, states are obsessed with their own security because they cannot rely upon a supranational body to guarantee it. Nor can nations hold 100 percent certainty about other nations’ motives. As a result, states will fear each other’s military capabilities and political intentions. They will take self-help measures and act strategically to guarantee their own survival, regardless of their internal politics.84 As Kenneth Waltz famously argued, a state’s place within the anarchical international system, rather than its domestic nature, will dictate national policy.85 Realists believe that international organizations and international law can do little to affect the pursuit of power and security. States may sign treaties, but the legal effect of the agreement does not affect their calculations. They will always pursue a course of conduct that maximizes their interests regardless of the presence or absence of a treaty. “Realists maintain that institutions are basically a reflection of the distribution of power in the world,” John Mearsheimer argued during the heady days of the end of the Cold War. “They are based on the self-interested calculations of the great powers, and they have no independent effect on state behavior.”86 Realism does not prevent nations from cooperating for mutual benefit. But the competition of states will make agreements difficult. First, because nations worry about their security from attack by other nations, they will be concerned by the relative gains of cooperation. Two nations might improve their situations in a deal, but if one of the two increases its relative advantage as a result, the balance of power might shift enough to sink the deal. In other words, states may refuse to cooperate, even if they were to enjoy absolute gains, if other nations gain even more. Second, nations may not cooperate even if they were to gain because they will worry that their partners will cheat. Without a world government, no higher authority can force sovereign states to keep their promises.87 These challenges make it unlikely that nations can commit to end the use of force in their relations. The 1928 Kellogg-Briand Pact, which purported to ban all war, stands as a failed symbol of such utopian efforts.88 Other forms of international cooperation may employ soaring rhetoric but—by design— produce little real world effect, such as the many resolutions of the UN General Assembly. But nations can cooperate in limited areas, so long that the resulting treaties or laws mirror the existing balance of power. Trade agreements, for example, allow comparative advantage to improve the economies of all signatories. Nations can cooperate to solve problems that cross borders, such as pollution or drug trafficking, or to divide valuable resources, such as underground oil fields or fisheries. States may even cooperate on matters of war if they can both benefit without either gaining a military or political advantage. State practice in waging armed conflicts produces customary practice that coalesces, over time, into the laws of war. Rules that will come to have the force of law must leave states better off in pursuing their interests. Nations will reject rules—no matter how noble their intent or humanitarian their goal—that leave them worse off. NATO allowed the United States and Western Europe to band together to contain the Soviet Union. NATO, however, did not alter the balance of power between Washington and Moscow. Instead, the institution served as a mechanism for the United States to better organize the forces on its side.89 NATO did not alter the balance of powers within Europe, but allowed the United States to rebuild Germany without reigniting security fears on the part of France and the United Kingdom.90 International agreements might also regulate wartime tactics, operations, and strategy. Nations at war, for example, might treat prisoners humanely if they can expect that their opponents will behave similarly.91 They might agree to foreswear chemical weapons if their use inflicts great suffering without giving either side an advantage.92 War’s high stakes, however, will tempt nations to cheat when battlefield conditions may make reneging harder to detect. It is also important to recognize that the success of a legal regime on war will depend on the international context. In a certain period of history, for example, a few great powers might protect their own security not just by building defenses, but also by maintaining a rough equality of power with their rivals. War might occur when nations band together to prevent a rising nation from upsetting the balance of power. In 1849, Great Britain and France dispatched troops across Europe to fight the Russian Empire in Crimea. Even though they had few interests there, London and Paris believed that Moscow’s move into Ottoman territory would upset the European balance of power. Using machine guns and trench warfare for the first time, the combatants fought to a stalemate. Under conditions that produced the static defenses of trench warfare, a balanced distribution of power and restrained national goals might make limitations on arms possible. No state would enjoy an advantage either before or after an agreement. It should be no surprise that the first treaties regulating warfare, the Hague Regulations of 1899, hail from the highpoint of Europe’s classic balance of power. Different circumstances, however, might not yield to noble goals. In the wake of World War I’s carnage, for example, the great powers sought to limit the naval arms race that they thought had caused British-German antagonism. If nations had built great fleets simply to keep up with their neighbors, a common limit on dreadnoughts might have restrained competition, thereby saving resources and reducing conflict. In the Washington Naval Treaty of 1920, the major western powers and Japan agreed to permanent limits on battleships, with the United States and Great Britain permitted to maintain the largest fleets, followed by Japan, France, and Italy.93 The Treaty of Versailles subjected Germany to far stricter limits on naval building.94 Wracked by the Great Depression and the rise of fascism, however, the Axis powers embarked on a campaign of territorial expansion. They circumvented the Washington and Versailles Treaties, either by concealing vessels (such as the Bismarck), building weapons that skirted the rules (Germany’s “pocket battleships”), or developing new naval weapons unforeseen by the drafters (aircraft carriers). As history suggests, changes in the international system and in war will have critical consequences for the nature and success of international rules. Several developments have set the environment that will surround the new military technologies. Perhaps the most important is that the destructiveness of war has rapidly declined since World War II. This seems counterintuitive in light of the deployment of vastly more destructive weapons by the superpowers and their allies. The United States used two fission bombs to destroy Hiroshima and Nagasaki, a level of destructive power that rivaled entire air campaigns of the day. A single modern fusion bomb would exceed the total destructive power of all of the conventional bombs dropped in World War II. Ballistic missile technology allows nations to deliver nuclear weapons anywhere in the world without having to deploy vast naval or air fleets. Throughout the 1980s, the United States and the Soviet Union fielded more than twenty thousand nuclear weapons each and thousands of air, sea, and ground missiles capable of delivering them. During the Cold War, both nations kept their nuclear forces at high levels of alert and could have destroyed the world many times over. But they did not. Though still present, conventional war has become less harmful. By some counts, the period from the end of World War II to the present day has seen the level of armed conflict between nations fall by an order of magnitude from the Peace of Westphalia to the mid-twentieth century.95 When corrected for the higher number of nations, wars between states have dropped both in their frequency and destructiveness. War no longer characterized the relations between European states. Indeed, for the first time in centuries, no major war between the great powers broke out in Europe or Asia. Historians now appreciate the Cold War as “the Long Peace.”96 But deaths from war have not disappeared. They have not even significantly declined. They have only dropped in wars between the great powers. While the number of conflicts between nations has steeply fallen, it has jumped inside states.97 Civil wars have replaced and even exceeded the scourge of great power war in terms of their casualties and frequency. By some estimates, post-World War II conflicts have killed forty million, including both combatants and civilians who have died on the battlefield or from related starvation and disease.98 Studies report that between 70 and 80 percent of these casualties occurred in civil wars.99 Conflict has become less global and more local. Globe-spanning wars between broad alliances of great powers, such as the Allies against the Axis in World War II, have receded. No great power has directly fought a war with another for the last seventy years. Instead, wars have remained limited to specific areas, such as the Balkans or the Middle East, with Africa generating the greatest share of conflicts and deaths. Realists account for the decline in war in two ways. First, and most important, the postwar world soon divided into a contest between two superpowers. Despite the arms race and proxy wars in Korea, Vietnam, and Afghanistan, this “bipolar” system had the counterintuitive effect of producing global stability. With only two superpowers, calculations of war and peace became simpler, less friction occurred, and the interest in superpower deterrence suppressed national interests that might have once caused war. According to this view, the multipolarity between the seventeenth and twentieth centuries made war more likely because of the heightened opportunities for conflict between more great powers. Second, the emergence of nuclear weapons only reinforced the balancing effect of bipolarity. The great powers could achieve greater security with a nuclear deterrent and the possibility of a nuclear exchange caused them to exercise greater caution in the use of force. The United States and the Soviet Union did not come into direct military conflict, it is argued, because of the fear that conventional war could escalate quickly into a nuclear conflict with devastation for both sides.100 As Waltz observed near the end of the Cold War, “[t]he probability of major war among states having nuclear weapons approaches zero.”101 The decline in interstate wars has accompanied greater international cooperation on nonsecurity areas, such as commerce, trade, and the environment. Nations have increased global welfare by lowering trade barriers within regional free-trade areas, such as the European Union and the North American Free Trade Agreement, and internationally through the World Trade Organization. These developments have encouraged scholars to see greater opportunities for international institutions and law to help build stability and peace. Such benefits usually arise from a hegemonic power, like Great Britain in the nineteenth century or the United States in the twentieth century, which can establish an international order built on free navigation and trade. Realist theory, however, would predict disorder when a dominant great power declines. Nevertheless, institutional scholars, as they are sometimes known in political science, argue that self-interested nations can cooperate to create international regimes that can maintain global order in the absence of a hegemon. In After Hegemony, for example, Robert Keohane argues, “Realist assumptions about world politics are consistent with the formation of institutionalized arrangements.”102 In other words, self-interested states that worried only about their security might still cooperate if they can realize benefits that leave them better off than before. There is little doubt that cooperation has advanced in the world economy even as the absolute and relative power of the United States has declined after the end of the Cold War.103 International agreements have successfully lowered tariffs and trade barriers and improved coordination in economics, science, environment, and health. Political economists theorize that these regimes help nations escape “the prisoner’s dilemma.” In this stylized game, the optimal outcome for two suspects is to remain silent under police questioning, but because they cannot communicate with each other there is a strong incentive to bargain for a reduced sentence and implicate the other criminal. The prisoner’s dilemma has become a shorthand in the social sciences for situations in which two parties, in pursuit of their rational self-interest, will do themselves short-term harm, when they could have achieved longer-term benefits if they had cooperated. Nations might engage in ruinous trade wars or arms races because of the prisoner’s dilemma. A classic example would be the nuclear arms race between the superpowers during the Cold War. Even though the United States and Russia apparently believe today that they need no more than 2,500 warheads each, during the 1950s and 1960s their nuclear stockpiles reached more than 25,000 weapons. Only decades of negotiation, verification, and trust-building deals on smaller weapons—like the 1988 intermediate-range missile ban—gave the superpowers the confidence to agree to reductions in their strategic arsenals. Liberal institutionalists argue that international agreements can help overcome the prisoner’s dilemma through repeated interaction. They believe that regimes offer states the chance to communicate and learn about each other over time in order to gain information and generate trust. As Andrew Guzman argues, a history of contact allows nations to develop reputations for keeping promises or retaliating against cheaters.104 Regimes can also improve the opportunities for cooperation by linking more issues together, which increase the benefits of performance and the tools for tit-for-tat responses to shirking. Permanent institutions can enhance these effects by spreading reliable information on the compliance of states and reducing transaction costs to future negotiations and deal making. Leading political scientists argue that the successful record in trade and finance should set an example to regulate other areas of global concern. There are several reasons to doubt, however, whether the structure of cooperation in areas such as the international economy will transfer easily to global security. First, and most important, cheating in a security agreement could inflict greater harms on a nation. If a nation suffers a surprise attack, loses a military advantage, or is left without its promised allies, it may encounter a long-term, permanent drop in its territory, population, resources, and ultimately power. Germany’s surprise 1941 invasion of the Soviet Union, and the breaking of their nonaggression pact, led to grievous losses and almost knocked Moscow out of the war. Withdrawal from a trade or financial agreement, by contrast, may cause economic pain, but likely cannot equal the severity and immediacy of a wartime defeat.105 Second, nations may have much greater concerns about “relative gains.” 106 In economic affairs, nations may remain content with gains in their GDP growth, increases in exports, and drops in the cost of imports, even if their trade partners do better. But in security affairs, as Joseph Grieco has argued, nations may refuse to cooperate if doing so would benefit a competitor in a relative sense.107 Cooperation is not impossible, but it is most likely in situations where both the use of force is not a serious threat between the nations concerned and they can engage in a long practice of tit-for-tat reciprocity to encourage cooperation.108 Cooperation itself does not disappear during armed conflict, but it is doubtful that formal legalization produces lasting rules of war. Nations have sought to follow some basic norms in combat, such as eliminating weapons that cause unnecessary human suffering. Despite the recent deterioration in practice, nation-states have generally refrained from using chemical weapons since the end of World War I. They have followed the Geneva Conventions on prisoners of war, though not consistently. Nations have observed others norms in the breach, such as the immunity of the civilian population and resources from attack. World War II included the aerial bombing of cities and the nuclear attacks on Japan, while the years since have seen precision targeting of terrorists off the battlefield, attacks on urban infrastructure, and the acceptance of high levels of collateral damage among civilians. International lawyers and diplomats may proclaim that nations follow universal rules on combat, either because of morality or a sense of legal obligations, but the record of practice tells a far different story. We should also not mistake temporary restraint in combat for a nations’ lack of military capacity. Human rights activists, for example, often hold up as an example of progress the 1998 Ottawa Convention banning land mines.109 Nongovernmental organizations (NGOs) led a decades-long campaign to persuade nations to give up antipersonnel mines, which can kill large numbers of civilians years after fighting has ended. In 1997, the Nobel Prize committee ultimately awarded the peace prize to Jody Williams, the director of the international campaign. The apparent success of the treaty, which now includes about 80 percent of the world’s nations, led to visions of a new approach to the laws of war in which groups and individuals, rather than states, would mobilize to force states to control war. Ottawa “spawned a new politics, new partnerships, new ways of thinking about the international environment. It was the forerunner of a clear notion of global citizenship,” writes Lloyd Axworthy, a former Canadian foreign minister and current university president. 110 “It challenged conventional notions of sovereignty and set in motion a form of coalition politics at the global level that could be used to shift power and political relationships.”111 NGOs and international activists have sought to duplicate the perceived success of the Ottawa Convention by extending its scope to regulate cluster munitions and small arms.112 Such efforts, however, unintentionally reaffirm the enduring importance of power. While the great majority of the world’s nations foreswear antipersonnel mines, most of them do not field large militaries or fight major conventional wars. Latin American states, for example, signed Ottawa in large numbers, but they have not fought a major war against each other for many decades. Nations with power projection capabilities, such as the United States, Russia, China, and India, did not commit to the agreement. Smaller states, such as Korea and Israel, faced with aggressive neighbors and local instability, also declined. Eliminating mines would place these nations at a serious disadvantage because they would reduce their ability to deter invasion or maintain a regional status quo. Despite their costs on civilians, mines actually may enhance stability and advance peace—they maintain borders by raising the cost of a ground invasion and have no offensive capability. Nations that live under threat of conflict will have little reason to agree to abandon such a defensive advantage that does not produce a corresponding reduction for offensive weapons. For similar reasons, the “Ottawa Process” will have little impact on other weapons. Nations with small militaries, peaceful neighbors, or large allies, such as many in Europe and Latin America, may promote new arms control agreements. But because the impact on their military performance is negligible and the chances that they will engage in war are small, these states experience very little loss in expected benefits—such as security gains by deploying these weapons—by signing on. Nations with large militaries and greater possibilities for conflict will not join agreements that could force a significant shift in their ability to prevail. Instead, nations will reach agreements whose reciprocal reductions in arms will leave no nation with an advantage. As James Morrow has observed, the laws of war treaties succeed when they guarantee minimum standards of treatment for captured soldiers that leave both sides to a conflict better off without giving either an advantage.113 A similar dynamic gives the agreements on chemical and biological weapons their strength: these weapons cause undue suffering and are difficult to control, and eliminating them leaves nations in the same relative power position as before. But if a nation can narrow a large gap in military power with a competitor by developing these weapons, it may violate an agreement banning their possession. Thus, nations at a disadvantage in conventional or nuclear forces have turned to biological and chemical weapons because reciprocity no longer holds. Nations, of course, will also follow humanitarian treaties even when they do not depend on reciprocity. But they will do so when compliance still advances their national interests. In the Korean and Vietnam Wars, for example, the United States followed the Geneva Conventions even though its opponents had not adopted the agreements. Rather than reciprocate, North Korea and the Viet Cong engaged in the systematic mistreatment of American prisoners of war. The United States, however, could have benefited from following Geneva anyway. Treating prisoners humanely may increase the willingness of enemy fighters to surrender, while a history of cruel treatment might cause opposing soldiers to fight to the death. These objectives may have even greater force if the contending armies are fighting for the support of the local populations. Extending Geneva protections to local guerrillas might help persuade the local populace to side with the government, for example. We should not overstate the influence of reputation on the decisions of nations at war. Human rights advocates, for example, might believe that nations should follow the Geneva Conventions to build a track record of compliance that will lead to better treatment for their soldiers in the next war. Under this argument, the United States should extend Geneva protections to al-Qaeda terrorists, even if the latter execute U.S. prisoners and civilians, because its soldiers will benefit in a future conflict with Iran or China. This claim, however, runs counter to the self-interest of nations at war. China would have little interest in punishing U.S. soldiers captured in a United States-China conflict because of American conduct toward terrorists in a separate contest. Such incentives might even cause a nation to conduct itself differently depending on the corresponding policies of its enemies. In World War II, for example, Germany generally followed the Geneva Conventions on the western front, where the United States and the United Kingdom treated prisoners of war well. The eastern front followed completely different rules: Germany and the Soviet Union descended into the barbaric treatment of each other’s troops and civilians. This detour into international humanitarian treaties here underscores the workings of reciprocity. Nations will follow the human rights treaties when they gain a benefit that outweighs the cost. Their captured troops will receive good treatment at the hands of the enemy even though Geneva will require them to bear costs in maintaining security, providing housing, and keeping prisoners in good health. Limitations on the use of weapons will follow a similar logic. Nations will refrain from using new weapons technologies only when they provide little benefit or their use by both sides leaves no one better off. Robots and cyberweapons do not bear the same features as the weapons where the laws of war have succeeded. They do not inflict unnecessary suffering out of proportion to their military advantages, as do poisoned bullets or blinding lasers. Rather, these weapons improve the precision of force such that they reduce human death and destruction in war. Nor do these new weapons technologies yet engage nations in a useless arms race. Nuclear weapons eventually became opportune for arms control because larger stockpiles provided marginal, if any, benefits due to the destructive potential of each weapon and the deterrence provided by even a modest arsenal. Mutual reductions could leave both sides in the same position as they were before the agreement. Today, the marginal cost of nuclear weapons for the United States and Russia so outweigh their marginal benefit that it is not even clear that a binding international agreement is needed to reduce their arsenals. Russia, for example, reduced its arsenal below New START’s ceilings of 1,550 nuclear warheads and 700 strategic launchers even before the United States approved the deal.114 Some experts estimate that the United States requires less than even those numbers to maintain an adequate deterrent, and it is possible that future presidents may reduce the nuclear arsenal without any international agreement.115 Cyberweapons and drones do not yet appear to bear these characteristics. The marginal gains in deploying these weapons will not be even across nations, but instead may be asymmetric. Some nations will experience much greater gains in military capability by developing cyber and drone technology. Or put differently, prohibition or regulation of these new weapons will not have equal impacts on rival nations. Chinese military doctrine, for example, emphasizes using such technologies to offset U.S. advantages in conventional sea and air power. 116 Drones might be used to overwhelm the defenses for U.S. carrier groups in the Western Pacific, while cyber weapons could interfere with U.S. command-and-control, communications, or even military readiness. Just as the Germans resorted to submarines to offset British and American naval superiority in World Wars I and II, the Chinese may well turn to robotics and network attacks in a future contest. Reciprocity will not hold because international limitations on new weapons would have the effect of favoring the United States’ existing advantage in conventional arms. Even if there might be some mutual advantage to a universal limit on these new weapons technologies, it is still doubtful that nations would accept a legal agreement. Nations would still develop these weapons because they improve their military capabilities at lower cost. Aerial drones, for example, can achieve the objective of attacking an enemy target at lower cost. An F-35 Lightning II stealth fighter/bomber costs $85 million under the best estimates, while Predator and Reaper drones cost about $4 million and $16.9 million respectively.117 The U.S. Air Force can not only purchase twenty Predators for the cost of a single F-35, but it can operate them at a far lower cost per hour, keep them on station for far longer (a predator can fly over a target area for up to fourteen hours) but also not risk the lives of pilots who may be captured or killed.118 Militaries will have an incentive to replace expensive manned aircraft with fleets of drones, regardless of whether their rivals choose to match them or not. Cyber and robotic weapons today resemble the emergence of the new weapons borne of the industrial age. At the time of their emergence, the impact of high-capacity firearms, long-range artillery, warplanes, submarines, and motorized vehicles was not well understood. World War I’s carnage can be attributed to the failure of nineteenth century tactics and strategy, still influenced by Napoleonic theories, to adapt to the greater firepower of modern weapons. Although nations and experts proposed the regulation of aerial and submarine warfare, their asymmetric benefits and the uncertainty of their effects on the balance of power precluded any agreement among the combatants. Nations may prove even more reluctant to adopt arms control schemes for weapons that have strong defensive qualities, such as the machine gun and artillery, because upgrading defensive systems may pose little offensive threat to neighbors.

Sign in to recut, update, or save this card.