Michael N. Schmitt
In an Articles of War post last week, Professor Terry Gill discussed his new book, The Use of Force and the International Legal System, co-authored with Dr. Kinga Tibori Szabó. It is a fascinating journey through the jus ad bellum, the law governing the use of force by States as an instrument of their national policy. Dancing elegantly across history, content, and controversies, the authors take a deep dive into the two pillars of the jus ad bellum, the prohibition on the “use of force” in Article 2(4) of the UN Charter and Article 51’s right of self- and collective-defense in the face of an “armed attack.” Both provisions are, in substantial part, reflective of customary international law.
Inspired by and building on their work, I would like to offer my thoughts on a topic the authors highlighted: the impact of new technologies of warfare on these rules. I begin by offering my perspective on how to approach that issue conceptually. I then turn to weapons characteristics that I believe have the potential to impact the jus ad bellum. In their study, Professor Gill and Dr. Szabó examined lethal autonomous weapons systems (LAWS), cyber and digital warfare, remotely piloted aerial vehicles (drones), and hypersonic weapons. I consider them and others to tease loose the features of certain weapons (a term I use here to include weapon systems and other means of directly harming an enemy, such as cyber capabilities) that potentially affect the content and effectiveness of this body of law.
Understanding Impact
In my view, weapons potentially affect the jus ad bellum in two ways. First, they can influence the actual content of the law, usually by causing States, tribunals, and experts to interpret aspects of the prohibition on the use of force and the right of self-defense in new or different ways. The paradigmatic example is the advent of the nuclear age, which eventually led in 1996 to the International Court of Justice’s (ICJ) Legality of the Threat or Use of Nuclear Weapons advisory opinion. Albeit controversial when issued, the opinion today is often treated as authoritative on the parameters of the jus ad bellum.
But new weapons can also spawn uncertainty as to the content of the law, the paradigmatic example in this regard being the advent of cyber operations in the late 20th century. Over a quarter century later, States are still struggling with how to apply international law in the cyber context, including the jus ad bellum issues of where the thresholds for qualifying cyber operations as a use of force or armed attack lie (Tallinn Manual 2.0, ch. 14 & 15).
There are many reasons to be concerned with such normative uncertainty. As I have discussed elsewhere, prominent among them is the risk of escalation, which is exacerbated when adversaries have different views on the threshold for particular response options. Consider a situation where one State conducts an operation believing it has stayed below the armed attack threshold that triggers the right of self-defense. But by its adversary’s understanding, the threshold has been crossed. The adversary accordingly resorts to force to defend itself. Believing the latter has mounted an armed attack against it, the first State responds forcibly, thereby launching an escalatory spiral.
A further concern is that when weapons create uncertainty as to the content of a jus ad bellum rule, some actors may exploit that uncertainty in anticipation of their adversary’s hesitancy to respond out of concern that its response may be viewed as unlawful. This is especially likely to occur when States that are not law-abiding face those that are.
The dynamic is, in fact, a fundamental quality of hybrid warfare. As Professor Gill and Dr. Szabó. note, “[m]uch of what is referred to as hybrid warfare takes place in a legal and factual ‘grey zone’” (p. 243). States will actively seek out this grey zone, as Russia does in cyberspace. For instance, although it is clear that cyber operations having significant physical consequences qualify as a use of force, the law is unclear regarding those not having such effects. This makes such operations appealing to a State that does not want to be seen as violating one of the core rules of international law, or that wishes to generate disagreement among other States about whether it has done so.
Therefore, one question explored in the following section is whether characteristics of some new weapons have caused, or will cause, us to interpret jus ad bellum rules differently than in the past or create uncertainty as to the content of the rules.
A second effect that weapons can exert on the jus ad bellum is an influence on “the degree to which the law can act as an effective restraint on their use” (Gill & Szabó, p. 238), in other words, deterrence. Law clearly has deterrent valence, for States do not like to be seen as acting unlawfully. This is evident, for instance, in Russia’s attempts to characterize its aggression against Ukraine in the language of self-defense and self-determination. Albeit nonsensical, the very fact of the claims illustrates the point.
Therefore, the second question in this approach is whether a characteristic of weapons makes it more or less likely that States will resort to force to advance or safeguard their national interests. Again, the paradigmatic example is nuclear weapons and concern about a devastating first strike against the target State’s own nuclear capacity (the so-called “use it or lose it” dilemma). That concern exerted significant pressure on States to act before their adversary. But once secure second-strike capabilities were developed (e.g., with ballistic missile submarines and missiles with multiple independently targetable reentry vehicles, or MIRVs), the opposite was true. Now, weapons exerted pressure not to act first because no decisive edge could be acquired.
Weapons and the Jus ad Bellum
Building on the work of Professor Gill and Dr. Szabó, I would like to highlight five characteristics of weapons that I believe have significant potential to affect either the content or the effectiveness of the jus ad bellum.
Complicating Defenses
The first is the ability of a weapon to make it difficult for an adversary to defend itself. There are two sides to this coin. On the one hand, States with such weapons are incentivized to strike first, to deliver a blow that will harm the adversary badly before it can act. On the other, States facing such weapons are incentivized to preempt their use by striking first. In this dynamic, the deterrent effect of the law is diminished. Because both sides may resort to force earlier than would otherwise be the case, the risk of an international dispute’s escalation through resort to the use of force rises.
Further, the harder it is to defend against a weapon, the greater the pressure to embrace more liberal interpretations of the law governing anticipatory self-defense, especially the notion of imminency. This has already occurred in response to the difficulty of defending against terrorism, as evidenced by the forward-leaning U.S. 2002 National Security Strategy released in the aftermath of 9/11. The logic of early preemption applies equally in situations where the adversary possesses weapons that are difficult to defend against.
There are numerous weapon characteristics capable of frustrating an adversary’s defense. For instance, speed can render a weapon difficult to defend against. Indeed, compressed reaction time was one of the reasons intercontinental ballistic missiles (ICBMs) and submarine-launched ballistic missiles (SLBMs) were (and remain) so problematic. The same challenge arises with cyber capabilities, which can generate effects almost instantly, and hypersonic weapons, which can reach a target before the adversary can effectively defend against them. Systems capable of great speed also facilitate attacking high-value targets in decapitation strikes or other tempting operations designed to frustrate enemy command and control as a conflict begins.
A weapon capable of surprise is likewise challenging to defend against. In this regard, stealth is particularly relevant. For instance, although cruise missiles are relatively slow, they are now being developed in stealthy (low-observable) variants (e.g., the air-launched Joint Air-to-Surface Standoff Missile, or JASSM), making them difficult to identify and therefore defend against. And it is possible to combine characteristics to exacerbate the adversary’s defensive challenge, as the U.S. Navy plans on doing in the near future by arming its stealthy Zumwalt Class destroyers with hypersonic missiles.
A further means of complicating an adversary’s defensive posture is to overwhelm it, which was the basis for developing a secure second-strike capability and MIRVing warheads. In contemporary warfare, the prospect of swarming drone attacks looms large. Drones are relatively cheap and easy to produce. The possibility of an adversary employing a large number of them (up to thousands) in a coordinated fashion is daunting.
And, of course, modern cyber capabilities are uniquely valuable for weakening enemy defenses. They can, for example, blind enemy early warning radar or disable defensive assets like surface-to-air missile systems to generate surprise or enable penetration. Cyber means can be used to insert misinformation into adversary systems to generate operational miscalculation. And they can render adversary decision-making and actions more transparent through presence inside the adversary’s systems.
It seems clear that the ability of these and other new weapons to neutralize or disrupt adversary defenses has the potential to weaken the restraining effect of the law and motivate reinterpretation of the self-defense’s imminency requirement.
Asymmetrical Capabilities
A weapon may provide an asymmetrical advantage that incentivizes attack (or preemption). Consider LAWS. Professor Gill and Dr. Szabó opined, correctly in my estimation, that the decision to engage in the use of force with the systems “will always be taken at the political/strategic level and not left to the weapons system” (p. 232). I agree.
However, weapons like LAWS or cyber capabilities may be so effective that use against an adversary not fielding them (or not having the ability to defend against them as discussed above) is irresistible. Putting law and morality aside, taking advantage of asymmetrical capability is logical from an operational perspective. Indeed, until the adversary fields comparable capabilities or effective countermeasures, a weapon may so alter the balance of power that it makes strategic sense to strike before the window of opportunity closes.
To offer but one example, LAWS or other autonomous systems can generate asymmetrical advantage by freeing up human resources for other tasks. Thus, LAWS could be used to channel enemy forces into kill zones maintained by humans who are on or in the loop in much the same ways that anti-personnel and anti-vehicle mines are used to channelize.
And cyber capabilities offer the prospect of launching crippling operations at a scale and speed not possible with kinetic operations and against assets and functions that would otherwise be out of reach. Notably, a State that otherwise is at an asymmetrical disadvantage may possess that capability. Although we have not seen such a campaign, the prospect of a societal “Pearl Harbor” is theoretically possible, including one against which defenses would be ineffective. A State wielding such capability may use it in the hope of delivering a devastating blow that quickly brings its adversary to its knees.
The prospect of such a cyber attack is already affecting the content of the law. Most significantly, States are increasingly open to interpreting the prohibition on using force, and even the right to use force in self- and collective defense against an armed attack, to include certain non-destructive and non-injurious cyber operations (e.g., France, Ireland, Singapore). In this regard, States (e.g., Germany) are embracing the approach proposed by the Tallinn Manual 2.0 experts (who maintained the first edition experts’ approach). It adopts the ICJ’s self-defense “scale and effects” test of the Paramilitary Activities (Nicaragua) judgment (para. 195) for use of force determinations (rule 69).
States are also implementing the Manual’s multi-factor assessment for determining whether the use of force and armed attack thresholds have been crossed (rule 69). The most prominent example is NATO’s embrace of both scale and effects and the multi-factors approach. Some States are even willing to extend one or both of these thresholds to cyber operations having widespread serious economic consequences, even though economic warfare had long been considered to fall outside the borders of the jus ad bellum (see, e.g., France, Denmark, and Norway)
Additionally, the fact that cyber operations place so many aspects of a society at risk has focused attention on the nature of their target in use of force and armed attack determinations. Target nature has always mattered in the jus ad bellum. For instance, actions directed against assets with a nexus to the State (e.g., military, diplomatic, or because it is flagged or registered in the State) are more likely to be characterized as unlawful uses of force or armed attacks than those where the nexus is less clear. But with the threat of hostile cyber operations multiplying, attention is increasingly being paid to the societal functions of the target. The growing propensity of States to single out cyber attacks against so-called “critical infrastructure” (see, e.g., Australia, Estonia, France, Norway, Singapore, Switzerland) is illustrative.
These trends beg the question of whether reinterpretation of the use of force and armed attack thresholds will have a knock-on effect with respect to the qualification of non-cyber operations as either. Complicating matters is the fact that while States are increasingly issuing opinio juris on how the jus ad bellum governs cyber operations, there is a lack of publicly visible State practice. So, the jury is still out on how cyber capabilities will eventually affect the interpretation and application of the rules.
Asymmetrical Risk
Weapons characteristics can also create a situation in which one side can act with relative impunity because the direct risk it faces is low relative to that of its adversary. This is most likely when it fields uncrewed systems. Because no attacker lives are at stake, that State may be more willing to engage in a use of force using a drone or LAWS than would otherwise be the case, as demonstrated by the preference for uncrewed instead of crewed aircraft when conducting cross-border, counter-terrorism attacks.
But there is another side to this coin. Uncrewed systems could encourage the use of force by the adversary. A State may be more willing to use force against such systems, knowing that their damage or destruction will be less likely to evoke a forcible response, as illustrated by Russia’s downing of a U.S. MQ-9 reconnaissance drone last year over the Black Sea or Iran’s shoot down of a U.S. MQ-4 over the Strait of Hormuz in 2019. This is so both because the harm to the other side does not involve the loss of human life and because of uncertainty about when uses of force, such as the shoot-down of an uncrewed platform, rise to the level of an armed attack triggering the right to a forcible response.
Weapon range is another factor that can yield asymmetrical advantage in terms of risk. The United States, for instance, can attack targets in most countries with long-range drones controlled from the United States or other countries far from the battlefield. It also exploits range by leveraging its naval assets and long-range aircraft. Few countries can strike back effectively or place the assets used at risk. And precision capabilities render it possible to use the platforms to hit targets with a low risk of collateral damage. This makes it more palatable to launch a strike by lowering the risk of domestic and international condemnation.
Complicating Attribution
To exercise the right of self-defense, the victim State must be able to legally or factually attribute the armed attack. Certain weapons complicate doing so. For example, those with the attribute of speed or surprise can render it difficult for the target State to determine who is attacking, if only because the latter is preoccupied with handling the consequences of the attack.
Cyber capabilities present the most significant challenge in this regard. This may be because the target State lacks the technical wherewithal to attribute reliably, because the attacker intentionally obfuscates attribution by, for instance, conducting a false flag operation, or even because the target State may be unaware it is under attack.
Consider the Stuxnet attack against the Natanz nuclear facility that the majority of the Tallinn Manual 2.0 experts believed qualified as a use of force and, perhaps, an armed attack (p. 342). It is widely believed that the United States and Israel were behind the operation (Operation Olympic Games). But it had been underway for well over a year before being discovered, and even then, there were many theories as to who had conducted it. By the time the spotlight turned on the United States and Israel, the operation was long terminated, meaning that even if it did qualify as an armed attack, Iran no longer enjoyed a right of self-defense.
As that case illustrated, a State that can escape factual attribution by employing a particular weapon has a greater incentive to use it, not only because a successful covert operation avoids condemnation for unlawful conduct altogether but also because even if discovered, the target State may lack a legal basis for responding in self-defense or the incident may have faded from international concern.
This attribution challenge may also affect how we understand mistaken action in the jus ad bellum. For instance, there is an ongoing discussion over whether mistakenly taking forcible action in self-defense, for example, because of the adversary’s effective use of false flag tactics, is an internationally wrongful act. Some argue it is not, so long as the attribution and responsive action were reasonable in the attendant circumstances. Others suggest that the mistaken States still acted wrongfully and should, therefore, be liable, for instance, for reparations.
Logic cuts both ways. On the one hand, the right of self-defense should arguably accommodate the challenge of attribution by allowing reasonable, albeit mistaken, responses. Conversely, the risk of misattribution to an innocent State augurs against response until sufficiently reliable evidence identifies the attacker. In either event, attribution challenges are influencing the legal discourse on the matter.
Death by a Thousand Cuts
Finally, cyber capabilities have drawn attention to the issue of aggregation. Before cyber operations became prominent, this issue was generally limited to assessing whether the self-defense threshold had been crossed, that is, determining when a series of uses of force would, in their totality, rise to the level of an armed attack (this is not an issue by the U.S. approach, which equates the armed attack threshold with that of the use of force, see Department of Defense Law of War Manual, §1.11.5.2). The use of force threshold was far less of an issue with kinetic operations because the physical consequences they caused usually crossed it.
In most cases, cyber operations cause non-physical consequences. Nevertheless, they lend themselves to a broad campaign of many related cyber operations that, in their entirety, could be devastatingly disruptive. This prospect has motivated States to consider whether consequences may be aggregated to reach the use of force and armed attack thresholds.
The trend is that they may. In its 2021 Brussels Summit Communiqué, NATO acknowledged this possibility by announcing that the “Allies recognise that the impact of significant malicious cumulative cyber activities might, in certain circumstances, be considered as amounting to an armed attack.” Individual States like France and Singapore have done likewise. And because all armed attacks are uses of force as a matter of law, the logic applies equally to aggregating cyber operations to reach the use of force threshold. Again, weapons characteristics are influencing State understanding of the jus ad bellum.
Concluding Thoughts
I compliment Professor Gill and Dr. Szabó for raising the issue of how weapons can influence the jus ad bellum. As explained above, that influence can manifest itself in shifting understandings of the content of the law or by affecting the valence of its rules in restraining uses of force. The takeaway lesson is that there is more to a legal assessment of new weaponry than a classic weapons review in accordance with Article 36 of Additional Protocol I for parties to that instrument or customary law for other States, like the United States. States considering the development and acquisition of weaponry, or watching an adversary launch a weapons program, should, therefore, bear in mind the broader question of their impact, for better or worse, on the jus ad bellum.
No comments:
Post a Comment