Ukraine Symposium – Using Cellphones to Gather and Transmit Military Information, A Postscript

In a previous post, Major Casey Biggerstaff and I analyzed the legal consequences of Ukrainian civilians using the “ePPO” application (app) to report incoming Russian air strikes. Under Article 51(3) of the 1977 Additional Protocol I to the Geneva Conventions, which binds both Russia and Ukraine, civilians lose their international humanitarian law (IHL) protections from attack “for such time as they take a direct part in hostilities” (see also ICRC Customary IHL study rule 6, DoD Law of War Manual § 5.8, Additional Protocol II Article 13(3)). We concluded that while using the app made great operational sense, ePPO users sometimes qualify as direct participants in hostilities and lose those protections. That means Russia may lawfully target those who do. Moreover, during an attack on another legitimate military objective, Russian forces need not consider any incidental harm to ePPO users during the proportionality and feasible precautions assessments that IHL requires.

Since publication, several IHL practitioners have contacted me regarding variations of the ePPO app scenario. Two merit a postscript to our piece: using an app to warn civilians of an impending attack; and crowdsourcing for intelligence reasons when the “crowd” is unsuspecting of the underlying purpose. As will be explained, in most cases, those involved cannot be characterized as directly participating in hostilities and would retain their various IHL protections from attacks.

Warning the Civilian Population

The first scenario involves Sentry, an app used in Syria to warn of Syrian and Russian air strikes. Developed by two Americans who had previously worked with Syrian civilians, it was a response to Russian and Syrian indiscriminate bombing, which they considered “the biggest threat to peace inside of Syria.”

Users of the app identify the type of aircraft that is spotted, what it is doing, and its direction. The Sentry program processes the data, estimates the aircraft’s trajectory, and transmits a warning through Facebook and Telegram messages, Tweets, and sirens, thereby enabling potentially affected civilians to take shelter. White Helmet rescue workers are also notified when and where the attack is expected to occur. If the area is populated, they inform medical facilities, which can then prepare for the arrival of the wounded and respond themselves.

The legal conundrum with such a system is that the military can also leverage the warnings, as in alerting air defenses or evacuating military facilities likely to be struck. This raises the question of whether the app’s civilian users are direct participants, a concept examined in greater depth in our earlier post. It suffices here to recall that it consists of three constitutive elements put forth by the ICRC in its Interpretive Guidance on the Notion of Direct Participation – threshold of harm, direct causation, and belligerent nexus. Although I take issue with some of the Interpretive Guidance’s application of the elements, I find them to be a reasonable interpretation of the direct participation rule.

The first element is evidently met. It requires that “[t]he act … be likely to adversely affect the military operations or military capacity of a party to an armed conflict or, alternatively, to inflict death, injury, or destruction on persons or objects protected against direct attack” (Interpretive Guidance at 46, see also ICRC Commentary to AP I ¶ 1944). Importantly, this only entails an effect on military operations.

Here, there is no question that the warning could deprive an attacker of the advantages it seeks. For instance, a strike on enemy forces in a building is foiled if they take shelter first because they have been warned. Similarly, the causal connection required by the second element is satisfied, for the relationship between the app’s use and defensive actions is plainly direct (Interpretive Guidance at 52). After all, the user’s report enables the military to act defensively within seconds of receiving the resulting warning.

It is upon the third element, belligerent nexus, that characterization of the reporting as direct participation falters. According to the Interpretive Guidance, “the act must be specifically designed to directly cause the required threshold of harm in support of a party to the conflict and to the detriment of another” (at 46). This does not mean that the actor’s subjective intent must be to harm one of the parties to the conflict. For instance, a civilian compelled to fight for an insurgent group is no less a direct participant; the civilian may not want to harm the enemy, but the act is such that it is meant to do so.

It is this “objective purpose” that is determinative. For example, the ePPO app we addressed in the previous post was designed to help the Ukrainian armed forces defend against attacks (see here and here). Because its objective purpose is to affect Russian military operations, the belligerent nexus element is undeniably satisfied. But here, the objective purpose of the Sentry app’s intended use is neither to adversely affect Syrian and Russian operations nor to enhance those of the Syrian opposition forces. Although it may have knock-on military consequences, it is meant to protect civilians. Using such an app is akin to telephoning civil defense authorities to enable effective attack warning and consequent sheltering. Surely, IHL does not allow a direct attack on someone telephoning civil defense authorities, even though an air raid alert to warn the population might enable military forces to defend themselves more effectively.

Furthermore, note that Additional Protocol I extends “special protection” to civil defense personnel. They must be “respected,” which means they are not subject to attack, and have to be allowed to perform their functions (art. 61(1)). This special protection extends to “civilians who, although not members of civilian civil defence organizations, respond to an appeal from the competent authorities and perform civil defence tasks under their control” (art. 61(2)). They lose that protection only if they engage in acts outside their civil defense duties that are “harmful” to the enemy (art. 65(1)). And even if a government has not expressly made such an appeal, the “object and purpose” of these and related rules is to help protect the civilian population from attack and facilitate an effective response following one. It would be incongruent to suggest that use of an app like Sentry would nevertheless qualify as direct participation; such an interpretation would fly in the face of that object and purpose.

I would caution that an act may have multiple objective purposes. For example, the purpose of swinging a baseball bat is usually to hit a ball. But it also is capable of being used to harm someone. If a civilian uses a bat to attack an enemy soldier, the objective purpose of the bat in the way it is used is to cause human injury. The same result would hold when a civilian misuses a Sentry-like app to warn the military of an attack. The app is capable of affecting military operations, and that is how it is being used in this situation. The individual is directly participating in hostilities.

Of course, the issue of doubt is always present when dealing with intent or purpose, whether subjective or objective. If there is a question as to the objective purpose of an act, the determination that it qualifies as direct participation must be made “in good faith based on information available to them in light of the circumstances ruling at the time” (DoD Law of War Manual § 5.4.3.2). But a degree of “doubt” about the objective purpose may still persist.

In that regard, both sides in the Russia-Ukraine conflict are bound by Additional Protocol I Article 50(1)’s rule that “[i]n case of doubt whether a person is a civilian, that person shall be considered to be a civilian,” a rule that the DoD Law of War Manual (wrongly) argues lacks customary status (§ 5.4.3.2). This presumption operates whenever the question is whether an individual is targetable, as it is in the case of direct participation.

But the presumption is, in my view, rebuttable. Lieutenant Colonel Michael Schauss and I have dealt with the issue of doubt elsewhere at great length. We suggest that the degree of doubt regarding status as a targetable person that operates to preclude an attack as a matter of law is context specific. The determination necessitates consideration of a complex array of factors ranging from the anticipated military advantage to the extent of uncertainty.

However, in the case of the apps, there usually would be little doubt with which to deal. While the ePPO app discussed in the earlier post is designed for belligerent purposes, the type at issue here is not. Only in cases where misuse of the latter to alert military forces is suspected would the issue of doubt as to objective purpose arise.

The Unwitting Participant

The second scenario is suggested by a Wall Street Journal report that a private company is paying users of an app to complete basic tasks such as snapping photographs, completing surveys, or reporting the price of food. The data collected is generally provided to private businesses seeking commercial information.

But the company, Premise Data Corp., is now also doing contract work with the United States and other governments using its network of (as of 2019) 600,000 contributors in 43 countries, including Iraq, Afghanistan, Syria, and Yemen. Some of the work is innocuous, as with a project to understand vaccine hesitancy. Yet, the Wall Street Journal also reported that

[i]n one pitch on its technology, prepared in 2019 for Combined Joint Special Operations Task Force-Afghanistan, Premise proposed three potential uses that could be carried out in a way that is ‘responsive to commander’s information requirements’: gauge the effectiveness of U.S. information operations; scout and map out key social structures such as mosques, banks and internet cafes; and covertly monitor cell-tower and Wi-Fi signals in a 100-square-kilometer area.

And according to the Wall Street Journal,

[d]ata from Wi-Fi networks, cell towers and mobile devices can be valuable to the military for situational awareness, target tracking and other intelligence purposes. There is also tracking potential in having a distributed network of phones acting as sensors, and knowing the signal strength of nearby cell towers and Wi-Fi access points can be useful when trying to jam communications during military operations. Nearby wireless-network names can also help identify where a device is, even if the GPS is off, communications experts say.

As the examples illustrate, data collection could lead directly to consequences reaching the direct participation threshold of harm. Therefore, whether the contributors are participating directly in hostilities depends on whether the belligerent nexus criterion is satisfied. And here, the distinction between subjective intent and objective purpose is critical, for, according to the report, the “tasks needed to be designed to ‘safeguard true intent’—meaning contributors wouldn’t necessarily be aware they were participating in a government operation.”

One could argue that even though the contributors harbor no intention of providing data for military purposes, the act (use of the app) has no intended purpose other than enhancing the military operations of the side receiving it. The fact that an individual contributor subjectively lacks hostile intent is legally irrelevant. It is the act that satisfies the objective purpose criterion.

The problem is that in other situations, IHL protects individuals whose actions affect military operations despite the objective purpose of the acts concerned. Consider human shields. The prevailing view (the correct one) is that voluntary human shields must be distinguished from involuntary ones; the latter are not subject to attack, and expected harm to them factors into proportionality and precautions assessments. This is so even though their presence undoubtedly affects the enemy’s willingness to attack. In extreme cases, their “act” of being present could so affect the proportionality calculation that an attack would be unlawful.

Or consider civilians ordered to evacuate a city in order to block an enemy advance. Although the underlying purpose of the evacuation is to affect enemy military operations, no reasonable IHL practitioner would suggest they are now subject to attack because the objective purpose of their presence is to block the enemy’s advance. Whether the civilians concerned know or are ignorant of how their movement is being exploited does not matter; they retain IHL’s protections from attack that they otherwise enjoy as civilians.

Here, the Interpretive Guidance advances a sound explanation for the apparent incongruity of results. It asserts there are exceptional circumstances where subjective intent matters. Unlike the cases of a civilian forced to engage in combat, here the civilians are “completely deprived of their physical freedom of action” (involuntary human shields, civilians forced to evacuate) or “totally unaware of the role they are playing in the conduct of hostilities” (unwitting human shields or civilians who evacuate based on intentionally false premises). As the Interpretive Guidance explains, “[c]ivilians in such extreme circumstances cannot be regarded as performing an action (i.e. as doing something) in any meaningful sense and, therefore, remain protected against direct attack despite the belligerent nexus of the military operation in which they are being instrumentalized” (at 60).

By this logic, civilians engaging in data collection who know or reasonably should know the objective purpose of their actions qualify as direct participants (assuming satisfaction of the first two elements). Those who have no reason to know do not.

Of course, in such cases, the doubt question may loom large because assessing subjective intent (what a person is thinking) can prove challenging. Be that as it may, I nevertheless would take the same approach to resolving that doubt as discussed above.

Conclusion

What distinguishes these two scenarios from that which Major Biggerstaff and I deconstructed in the previous post is the fact that the objective purpose of the ePPO application was to impede Russian operations. Moreover, its users were aware of that purpose and intended to fulfill it. By contrast, civilians who engage in activities that do not have an objective purpose of affecting military operations or capabilities are not direct participants in hostilities. This is so even if their actions have that effect. Thus, employing an app to enhance civilian protection from attacks does not expose its users to loss of IHL’s protection. And in exceptional circumstances in which users are unaware that an app is being used for military ends, they likewise retain that protection.

***

Michael N. Schmitt is the G. Norman Lieber Distinguished Scholar at the United States Military Academy at West Point. He is also Professor of Public International Law at the University of Reading; Professor Emeritus and Charles H. Stockton Distinguished Scholar-in-Residence at the United States Naval War College; and Strauss Center Distinguished Scholar and Visiting Professor of Law at the University of Texas.

Photo credit: Unsplash