By: Brian Lee Cox

Editorial note: This essay is part of a cross-platform online collaboration with Just Security, Lawfire, and Opinio Juris in recognition of the fifth anniversary of the attack on the MSF trauma center in Kunduz, Afghanistan. For the other articles from the online collaboration, please see the links below.

---

Exactly five years have passed since the horrific attack by the U.S. military on the Médecins Sans Frontières (“MSF”) trauma center in Kunduz, Afghanistan, and the redacted official military investigation was released to the public more than four years ago. On the day the investigation report was released to the public, MSF published a list of ten questions posed to the U.S. military in regards to the official investigation. Now, more than four years later, no official response to these questions has been made publicly available.

The questions appeal to matters with enduring relevance to the study and practice of the law involving armed conflict, such as protections for medical workers and humanitarian activities during war and the pursuit of effective accountability following attacks on civilians in armed conflict. The organization deserves meaningful answers to the unresolved questions, and informed answers can provide a useful contribution to the study and practice of international law and military justice.

While I am no longer a member of the U.S. military and my answers thus do not qualify as an official response, based on my professional endeavors in my capacity as a U.S. Army judge advocate I am deeply familiar with the circumstances of the attack and the official response following the airstrike. I have engaged in extensive conversations regarding the airstrike with many friends and former colleagues who were directly involved with advising various echelons of military command after the tragic attack, and, before I retired from military service in 2018, I was assigned to develop in-depth training for fellow judge advocates involving the Kunduz airstrike and lessons learned following the attack.

My study of the airstrike and the official and public narratives that ensued has continued since my retirement, and I am currently completing a project that engages in a thorough critical assessment of the official investigation and lessons “learned” therefrom. While that project is nearing completion, the fifth anniversary of the attack on the MSF trauma center in Kunduz, Afghanistan presents an opportunity to address, from an informed insider’s perspective, the questions posed by MSF as an organization to the U.S. military.

MSF as an institution deserves answers to these unresolved questions, but official answers are not on offer. The surviving victims of the horrendous attack, the families of those killed, and the organization that continues to provide vital medical services across the globe in the most dangerous circumstances imaginable, deserve better clarity than is currently available from official sources regarding what went wrong with the attack and what was done about it.

This imperative inspired the current cross-platform collaboration on the fifth anniversary of the airstrike. A central aspect of the ongoing discussion of the attack and its broader implications is providing the answers that the victims, families, and MSF association deserve. This essay offers an insider’s perspective regarding the answers to the unresolved questions posed by MSF. The essay is structured around those ten questions by copying them from the applicable MSF webpage and answering each in turn.

1. What was the physical description of the intended target provided by the Afghan forces and how did it match the description of the MSF hospital?

The official investigation report does not contain direct evidence of the physical description of the intended target that was provided by Afghan forces to the U.S. Special Forces (“USSF”) element. The description was provided over a telephone and through an interpreter in person to the U.S. ground force commander that was co-located at the provincial governor’s compound with some members of the Afghan ground element. However, the available evidence is consistent regarding the nature of the description and how it was communicated.

This is the sequence described by the ground force commander when the aircrew of the supporting AC-130 gunship requested a physical description of the proposed target after the target acquisition and identification system aboard the aircraft pointed to “the center of a field” once the crew entered the grid coordinates provided by the ground element to the aircrew: “I asked the [redacted, likely from the context a reference to a member of the Afghan ground element] and the description I got, and this is before they even SP’d [‘starting point,’ a colloquial method for describing initiation of a movement] from [redacted, likely from the context a reference to Camp Pamir at the airfield to the south of Kunduz City] was a long T-shaped building with a small off shoot. I can’t remember the word I would have used for it. In a walled in compound with multiple out buildings and there was a gate facing to the north with an arch” (pg. 389). Whatever the precise description provided by the joint terminal attack controller (“JTAC”) to the aircrew, the aircrew focused on the “T-shaped building” portion of the description and identified the main building of the trauma center about 400 meters south of the point on the ground initially indicated by the aircraft target acquisition and identification system.

Based on the circumstances at the base of operations at the time, further clarification of the events is unlikely. As the ground force commander recalls during his interview with the investigation team, “If I wanted to get information, Sir, I would have to talk to my interpreter, who would talk to the [redacted, likely a reference to the specific unit of the Afghan Special Security Forces (‘ASSF’) involved in the operation] who would call to somebody” (pg. 387). The ground force commander indicated that he could not recall the name of the member of the ASSF with whom he communicated (pg. 387), and he also does not know with whom this point of contact was communicating over the telephone (pg. 388).

The assertion that Afghan forces may have intentionally provided a physical description that matched the MSF trauma center in hopes that the U.S. military would attack the hospital is addressed briefly elsewhere in discourse involving the attack. This is a topic I address at length in my full study of the official investigation, but a summary is in order here.  While I do not completely rule out the possibility based on my study of the available evidence, my conclusion is that this allegation is incredibly unlikely.

The transcript of the interview with the ground force commander does indicate that the ASSF ground assault force that was planning to move back into the city from Camp Pamir did intend to conduct raids on two different targets—one was described as an “National Directorate of Security (‘NDS’) prison” and the other was not described. This creates the possibility that the second target was the MSF trauma center, but the ground force commander recalls that the ASSF members “just said it was further north, like up around the market somewhere” (pg. 387). The MSF trauma center is located to the south of the NDS facility, not “further north.” Furthermore, if this is a reference to the main market in the center of Kunduz City, this is 600 meters or so north of the compound of the provincial governor where U.S. and Afghan forces had established their base of operations. The spatial orientation of the various points of interest alone makes it factually unlikely, wholly apart from the issue of potential motivation and likely consequences, that Afghan forces intentionally duped the U.S. military into targeting the MSF trauma center by deliberately providing a target description that matches the hospital.

2. Which forces—either U.S. or Afghan—on the ground had “eyes on the target” [National Directorate of Security compound] before and during the attack? And if the U.S. forces did not, why not?

Neither USSF nor ASSF ground elements had “eyes on” the intended target. The security situation at the base of operations Afghan and U.S. military forces established in Kunduz City would have prevented the group from sending anyone out of the compound to establish eyes on. The ground force commander recalls that on the day of the airstrike, the group experienced “the heaviest contact we had received up to that point” in the mission (pg. 530). The group was rationing battery power to maintain some limited radio contact with higher headquarters and supporting aircraft, and sending a contingent out of the secured base of operations—referred to alternately as the PGOV (“provincial governor”) or PCOP (“provincial chief of police”) compound—in search of the NDS facility would not have been tactically feasible.

Depending on the particular targeting circumstances and the specific aircraft being utilized, a lack of visual contact with the target by ground forces can significantly increase the risk of target misidentification. However, the loitering overhead orbit of the AC-130 and the sensor suite aboard the aircraft actually put the aircrew of this particular aircraft in a much better position to identify the target than the ground element.

The visual field available for crews of most fixed wing aircraft is often described as looking through a “soda straw” when trying to identify targets on the ground. The sensor suite on an AC-130, in contrast, includes a multispectral television sensor and a high definition infrared sensor, both of which were utilized extensively in the target identification process. As examined below, the central factors that led to the tragic outcome of the Kunduz airstrike were inadvertent application of existing ambiguities in doctrinal terminology and a resulting lack of clarity involving the intent of the ground force commander.

The inability of the ground element to put eyes on the target was not a significant contributing factor. The aircrew observed the target area for a total of 68 minutes before initiating the airstrike (pg. 027) and, when they did, they attacked the target they intended to strike.

3. During the airstrike, how many calls or warnings were record[ed] from Afghan Special Forces to the U.S. Special Forces commander or the Joint Terminal Attack Controller to inform them of the targeting error, i.e. that the intended target (National Directorate of Security compound) was not being hit?

There is no indication from the factual record that members of the ASSF element were aware of the targeting error. At the time that the AC-130 initiated the strike, the so-called “battle damage assessment (‘BDA’) video” (colloquially referred to as the “gun tape”) indicates that the ASSF ground assault force (“GAF”) convoy was five kilometers away from the (mis)identified target (pg. 066, footnote 179).

The multispectral television sensor operator on the AC-130 recalls in a sworn statement provided to the investigation team after the airstrike that the GAF was “stopped around 650m from the Obj” [objective] even after the airstrike and that the convoy “didn’t advanced (sic) to the compound before we had to RTB” (return to base) to refuel (pg. 363).

The ASSF ground element was not physically in a position, then, to determine exactly which target was being attacked. The members of the U.S. ground element at the PCOP compound were not aware that the wrong target had been attacked until after the airstrike concluded, which is further evidence indicating that they did not receive any calls or warnings from the ASSF about the targeting error during the attack.

4. Why was the attack not called off before the AC-130 completed its mission? Who would have been responsible to call off the attack of an AC-130 deprived of its essential communication capacity operating in a densely populated area? And why didn’t he/she?

The airstrike was not called off before the AC-130 completed its mission because the personnel directly involved in the attack remained under the mistaken belief that the targeted compound was a “Taliban stronghold” for the duration of the attack. Because a “self-defense” rules of engagement (“ROE”) justification was provided by the JTAC to the aircrew, approval from a higher echelon of command was not required.

The inadvertent misapplication of applicable rules of engagement is discussed in more detail below in the response to Question #6. For present purposes, though, this explains why the approval chain for the attack did not extend beyond the ground force commander. The ground force commander (“GFC”) was the “on-scene commander” in doctrinal terminology, and in ROE terms commanders “always retain the inherent right” of self-defense.

Beyond the personnel directly involved in the attack, the chief of operations, or “CHOPS” as the position is colloquially called, at the Special Operations Task Force-Afghanistan (“SOTF-A”) headquarters element was in all likelihood the person at the lowest level of responsibility that could have stopped the attack if the targeting mistake had been realized in time. The CHOPS leads the current operations integration cell (“COIC”), and the COIC, in turn, is “the focal point for controlling the execution of operations.” Because the SOTF-A headquarters was the lowest level of command on the task organization chart (pg. 040) above the ground force commander (as the “commander” of “Area of Operations-North”), the SOTF-A CHOPS likely would have been best situated to intervene and stop the airstrike if the targeting mistake had been recognized in time.

When the initial call was made by the MSF Country Director at 2:19 AM, which was 11 minutes after the airstrike began, the headquarters member on the receiving end of the call went to the Joint Operations Center (“JOC”)—a term more widely used than COIC—to ask about an airstrike. Because of the prevailing operational circumstances, no one in the JOC realized at that time that it was the AC-130 that was in fact targeting the trauma center (pg. 067–068).

The person with whom the MSF Country Director was communicating eventually realized that it must be the AC-130 that was attacking the trauma center. This person then entered the JOC and declared, at 2:33 AM, “You’re hitting the Trauma Center” (pg. 068). Notwithstanding this revelation, the attack continued for another four minutes, until 2:37 AM, according to the investigation report.

The report does include as an exhibit a sworn statement from the person I assess, based on the context of the statement, to be the SOTF-A CHOPS. In this statement, the (presumably) CHOPS recalls, likely referring to the point at which the revelation was made that the trauma center was under attack, “My initial reaction was to do nothing about the strike because I did not know the situation on the ground that warranted the strike to be conducted, and as the [redacted] I will always give the [redacted, presumably a reference to the GFC] the benefit of the doubt when conducting engagements in self-defense” (pg. 435). At 2:38 AM, the (presumably) SOTF-A CHOPS called for a ceasefire, but the last round had already been fired at the trauma center one minute earlier.

This mosaic assemblage of circumstances accumulated from throughout the full investigation report describes the personnel who would have been responsible for calling off the attack if they had been sufficiently certain that the MSF trauma center was being attacked. Regarding the aspect of the question involving deprivation of essential communication capacity, technological communications challenges existed but did not constitute a significant factor in the targeting error. As unfortunate as it is, military operations in densely populated areas have become the norm rather than the exception in contemporary warfare. This phenomenon has invariably increased the risk to civilians, but has not obviated the necessity to use force in densely populated areas.

5. How do you reconcile the difference in the duration of the attack between the U.S. version of events and the MSF internal review?

I currently have no adequate explanation for the difference, but I do note that the investigation team had access to detailed transcripts of radio communications, video from the “battle damage assessment recorder” of the AC-130 (pg. 062, among others), and a log of electronic messages sent from the SOTF-A JOC to personnel with the ground element at the PCOP compound (pg. 512). After reviewing information from these and other sources, the investigation report concludes that the attack lasted for “30 minutes, 8 seconds” (pg. 027).

The official account differs significantly from the findings published in the MSF internal review, which estimates that the attack lasted approximately an hour or more and that it ended as late as 3:15 AM. While I do not have access to the underlying data used to develop this estimate, I would welcome the opportunity to review the data in an effort to reconcile the difference between the accounts of the duration of the attack.

6. Given that U.S. troops were not directly under fire, what are the limits of engagement of U.S forces in combat activities?

The answer to this question gets to the very heart of what truly went wrong with the attack. General Joseph Votel, who was commander of U.S. Central Command when the investigation was released to the public and who briefed MSF representatives before he announced the release, was asked by a reporter at the press conference announcing the release of the investigation if there is “something in particular with the human error . . . that jumps out at you?”

There absolutely is one factor “in particular” that led to the tragic outcome of the airstrike, but the “human error” exists far beyond the personnel involved in the attack. To date, this human error has not been identified by the U.S. military, and it existed long before the Kunduz airstrike and persists now even five years on. General Votel’s answer, that “the communication between the air and the ground” merited “more discussion between what was going on” is well short of the mark, but the same is true for the wider institutional effort to identify what truly went wrong.

The single factor that contributed most to the “human error” involved in the attack on the MSF hospital involves application of ambiguous language present in existing rules of engagement and other use of force policy. In this particular attack, the ambiguity responsible for the human error involved comes down to the use of one word: hostile.

A full assessment of the functional effect of this and other ambiguities in existing practice and doctrine is not feasible in the current format, but for this essay I can summarize the findings of my comprehensive evaluation of the official investigation and lessons “learned” therefrom. The depiction that emerges directly addresses the question of the “limits of engagement” in the Kunduz airstrike “given that U.S. troops were not directly under fire.”

There is no provision of the rules of engagement that categorically prohibits attacking a target even if troops are “not directly under fire,” but the approval process for such an attack is different than for a so-called “self-defense” engagement. The trouble is, the same word is used doctrinally to describe both an “offensive” and a “self-defense” rules of engagement authorization. The rules of engagement permit the use of force in self-defense in response to a “hostile act” or demonstration of “hostile intent,” while describing an “offensive” engagement that is not in response to a perceived hostile act or demonstration of hostile intent as an attack against a “declared hostile force.”

This ambiguity alone is problematic, but a there is a third doctrinal use of the word “hostile” that, in effect, creates a bridge between a “self-defense” and an “offensive” use of force. Both colloquially and doctrinally, the word “hostile” is also used more generally to describe an adversary without regard to the potential ROE characterization of the adversary.

Examples of this imprecise use of the word “hostile” in military doctrine and practice are too abundant to count, but the U.S. Department of Defense Law of War Manual observes, as but one example, that “spying, sabotage, and other hostileactivities behind enemy lines would render a person liable to prosecution for such conduct” (emphasis added).  Another doctrinal example, this from a publication focused on protections for civilians in armed conflict, warns that “in some cases hostile actors may attempt to blend in with” displaced persons and separately suggests to military planners that training exercises should “include civilians who are not hostile” (emphasis added).

While this ambiguity regarding use of the word “hostile” may seem rather innocuous, it manifests in targeting mishaps with devastating consequences. This ambiguity is the single most significant factor that led to the tragic outcome of the Kunduz airstrike.

The “hostile” trail begins with the description of the intent for the close air support that was communicated by the ground force commander to the JTAC. As the ground force commander recalls in the interview with the investigation team after the attack, he remembers “very specifically spelling” out his intent for the requested close air support “to reduce heavy weapons and strong points to allow” the ASSF ground assault force to maneuver to their objective, the Afghan government NDS facility (pg. 391). The ground force commander further clarifies that he wanted to identify the target “prior to the actual contact” in order to “make sure that we were on target and that everyone understood exactly what needed to happen” before the ASSF ground assault force became decisively engaged en route to the objective (pg. 391).

This intent is consistent purely with the “self-defense” usage of the word “hostile,” which permits a use of force in response to a “hostile” act or demonstration of “hostile” intent. The JTAC summarizes this intent in an ambiguous manner that could be consistent with either a “self-defense” or “offensive” engagement. Later, as the aircrew is trying to identify the intended target and describes the activity of the people they are observing at what was in fact the trauma center, the JTAC asserts that the “compound is currently under the control of the TB [Taliban], so those 9 PAX [personnel] are hostile” (pg. 033, emphasis added). This meaning for the word “hostile” is consistent with the doctrinal usage to describe an adversary without regard to ROE classification.

While the ground force commander intended the requested close air support to be purely for the defense of the ASSF and the JTAC instead communicated the non-descript usage of the word “hostile,” the aircrew interpreted the intent in a manner consistent with the “offensive” ROE application. The transcript of a radio conversation between the navigator and fire control officer of the supporting AC-130 reveals that at one point the navigator believes that the JTAC has “already confirmed that this prison complex [factually the MSF trauma center] is hostile” (pg. 060). Although this usage, too, is consistent with the ROE ambivalent meaning of the word “hostile,” the recollections of the members of the aircrew that are reflected in their sworn statements after the attack reveal that they interpreted the observation by the JTAC in a manner consistent with an “offensive” use of force.

The TV sensor operator observes in a sworn statement, for example, that the JTAC “said [before the attack that] the compound was under enemy control and that those [redacted] were declared hostile” (pg. 363, emphasis added). In a sworn statement that appears from the context to be provided by the pilot of the AC-130, the affiant recalls that the JTAC “relayed to us that the complex that we were monitoring [factually the trauma center] was under ‘TB’ (Taliban) control and that all associated personnel were declared hostile” (pg. 492, emphasis added). Similarly, in a sworn statement that from the context appears to be provided by the electronic warfare officer aboard the AC-130, the affiant recalls that the JTAC determined the compound being observed by the aircrew “was under Taliban Control and declared all personnel in that compound to be hostile” (pg. 500, emphasis added).

The practical effect of interpreting “hostile” in this manner is that, in ROE terms, “Once a force is declared hostile by appropriate authority, U.S. forces need not observe a hostile act or demonstrated hostile intent before engaging the declared hostile force” (emphasis added). The JTAC, of course, is not an “appropriate authority” to make such a “hostile” declaration, and the official investigation report makes clear that no one involved in the attack was conferred with such authority.

Due to the widespread ambiguity in doctrine and practice related to the term “hostile,” however, no one involved in the attack knew the difference. The JTAC proffered a ROE justification that was consistent with a self-defense engagement, but also used the term “hostile” in the ROE-neutral meaning without being aware that the aircrew would interpret this to be declaring the people being observed at the MSF trauma center to be a “hostile force.” The aircrew accepted the “self-defense” ROE justification, which permits a use of force in response to a “hostile” act or demonstration of “hostile” intent, but unwittingly interpreted the ROE-neutral “hostile” observation as an assertion that the people being observed were a “declared hostile force.”

If this mistake had been recognized, the airstrike would have never taken place. That is, if the members of the aircrew were aware that the ground force commander’s actual intent was to identify the target before the ASSF ground assault force came under attack and then to “reduce heavy weapons and strongpoints” in defense of the GAF, the aircrew would not have initiated the attack when the JTAC directed them to do so. Due to several factual circumstances, the ground force commander believed the ASSF ground assault force to be under attack when he advised the JTAC to direct the AC-130 to initiate. He also believed the aircrew was operating in accordance with his intent to defend the ASSF GAF and that “there shouldn’t have been any confusion” after he had clarified his intent earlier (pg. 391).

In actuality, the aircrew observed that the ASSF ground assault force was several hundred meters from the perceived objective, the MSF trauma center, when the airstrike was initiated. The aircrew never visually identified any weapons at the compound they were observing, let alone “heavy weapons and strongpoints” as described by the ground force commander. The absence of weapons is to be expected since the members of the aircrew were in fact observing the MSF trauma center and the hospital had a strict “no weapons” policy. At the command to initiate the attack, then, the aircrew would have been fully aware that they were not carrying out the intent of the ground force commander if they would have properly understood the intent. This would have prompted the aircrew not to engage, and they never would have done so since the ASSF ground assault force never came under fire from the MSF trauma center.

It is the existing ambiguity regarding practical and doctrinal application of the term “hostile” that caused the fateful misunderstanding. This is without a doubt “human error” as the investigation and official narratives indicate. However, the error is not attributable to the humans directly involved in the attack.

In attacking without first observing a “hostile act” or demonstration of “hostile intent,” the aircrew unwittingly conducted an attack in a manner consistent with the ROE designation of a “declared hostile force” while still believing they were carrying out the intent of the ground force commander. The bridge from the ROE “self-defense” usage of the term “hostile” (intended by the ground force commander) to the ROE-neutral description of an adversary (communicated by the JTAC) to the ROE “offensive” usage to connote a “declared hostile force” (interpreted and applied by the aircrew) set the conditions for the aircrew to “violate” the ROE by engaging in an “offensive” attack without authorization. The “human error” existed long before the airstrike, and it remains unidentified in existing doctrine and training to this day.

What, then, are the “limits of engagement” when U.S. forces “in combat activities” are not “directly under fire”? The short answer to that question is that the U.S. forces involved unquestionably exceeded the existing limits. The trouble is that the personnel involved in the attack were not aware that they were exceeding the limits of engagement. As General Votel observed during the press conference announcing the release of the investigation to the public, the personnel “were trying to get to the right answer here and do the exact right thing.” Deficiencies in existing doctrine and practice prevented them from doing the “exact right thing,” but this central contributing factor has yet to be identified, let alone rectified, by the U.S. military.

This examination of the application of relevant ROE is not complete and comprehensive. The NATO version of the ROE was also involved in the mistaken attack, though the ambiguity pertaining to the term “hostile” is also present there. A more complete assessment of the role of both the U.S. and NATO sets of ROE is possible by drawing on the classified versions of both. However, most of the provisions that are relevant to the Kunduz airstrike are present in the versions of the ROE that are available to the public.

Finally, sources of use of force policy other than the ROE, such as the “Tactical Guidance,” implementation of the “no-strike list” database, “clearance of fires” procedures, doctrine related to attaining “positive identification,” among others, are addressed in my comprehensive assessment of the official investigation and the lessons “learned” therefrom. However, for present purposes, the abbreviated examination of the ROE that is focused on ambiguous application of the term “hostile” is sufficient to address the “failure” of U.S. forces to comply with the relevant “limits of engagement” when, as in the Kunduz airstrike, the forces are not “directly under fire.”

7. How do you justify that no court martial is recommended in response to the killing of 42 people and considering the protected status of the hospital?

As tragic and horrific as the outcome of the attack is, decisions regarding potential legal sanctions for the personnel involved in the attack must be made in consideration of the process that led to the airstrike. If the personnel had been aware of the protected status of the hospital and attacked it anyway, the outcome of the decision not to initiate court-martial charges would have been very different.

I have discussed the circumstances of the airstrike and the disposition decisions with many friends and former colleagues who were directly involved in advising the command teams that decided not to initiate charges. Without exception, these judge advocates have communicated to me that they are of the opinion that criminal charges were not appropriate based on the process that led to the strike.

The issue of perspectives regarding accountability for unintended outcomes for attacks in armed conflict is addressed further in the closing reflections of this post. And I consider the issue of accountability in greater detail in the Just Security and Lawfire posts of this collaboration. However, based on my detailed review of the investigation and other official accounts of the circumstances of the attack, I endorse the decision not to initiate court-martial proceedings.

This does not mean that I think accountability for the airstrike is not vitally important. But I worry that focusing on accountability for individual personnel based on the outcome of the airstrike risks diverting attention away from the institution that bears responsibility for the process that led to the tragic outcome.

8. Who within the chain of command is ultimately responsible for the deaths of 42 people, and how is that person being held accountable?

The matter of accountability is addressed in the answer to the previous question involving the decision not to initiate court-martial proceedings for anyone directly involved in or otherwise responsible for the attack. As I note below in concluding reflections, true “accountability” in the context of the attack on the MSF trauma center needs to be centered on the U.S. military as an institution rather than any individual person directly involved in or otherwise responsible for the attack.

9. Does the United States consider the actions of U.S. forces to be negligent, and if not, why?

The official report is replete with findings that would indicate that the personnel involved in the attack were negligent, probably even reckless, and potentially even worse. If the outcome of the attack is the focus of the assessment, it is not controversial to conclude that the personnel were at least negligent. However, as I describe in other posts, the official investigation is unfortunately an entirely unreliable source of information. Aside from the findings of the official investigation, if the process of the attack is the focus, a characterization of “negligence” is not as straightforward.

There is no doubt that the personnel involved in the attack got it horribly wrong. I conclude in the full assessment of the official investigation that the navigator and the fire control officer of the AC-130 aircrew did not comply with the requirement to take feasible precautions in the attack, and this can certainly be characterized as negligent.

My conclusions regarding these violations of the feasible precautions rule, however, are complicated by the fact that these conclusions in essence involve a failure to comply with the doctrinal duty to confirm “validity with respect to” the applicable ROE. These personnel should have clarified the uncertainty they seemed to have regarding application of the ROE, but ultimately all the personnel, including the navigator and the fire control officer, appeared to genuinely believe they were complying with applicable use of force law and policy.

My assessment is that the truly culpable negligence rests with the entire U.S. military as an enterprise. The official investigation report provides abundant evidence that the personnel involved in the attack on the MSF trauma center were thoroughly trained in relevant use of force law and policy. The investigation itself was purportedly conducted “with painstaking attention to detail” and was “followed by an in-depth review process.”

As the response to Question #6 describes, the single most significant factor in the tragic outcome of the attack was utilization of ambiguous terminology that was consistent with existing doctrine and that nonetheless led to a “violation” of the relevant ROE. The official investigation report not only fails to identify the true nature of the factors that contributed to the tragedy, but it systematically mischaracterizes even the most basic aspects of relevant use of force law and policy. That this report is purportedly the product of “painstaking attention to detail” that was “followed by an in-depth review process” further impugns the enterprise that generated such an absolute failure to account for what went wrong.

The discernible negligence here, then, rests with the enterprise that trained the personnel involved in the airstrike and those responsible for evaluating the attack afterward. As I observe below in the concluding reflections, this systemic negligence is correctable. However, it is the enterprise, rather than the products of that enterprise, that is the one truly deserving of the characterization of “negligent.”

10. How will the disciplinary measures chosen deter U.S. military personnel from violating the laws of war in the future?

In short, the disciplinary measures imposed will have very little effect in preventing the circumstances that caused the catastrophic outcome of the Kunduz airstrike from contributing to a similar tragic outcome again in the future. This is not because the administrative measures imposed are inadequate. Rather, the disciplinary measures that were imposed are misdirected.

Although the outcome of the airstrike is horrendous, the personnel involved, as General Votel asserts, “were trying to get to the right answer here and do the exact right thing.” The aircrew observed the target area for 68 minutes before initiating the airstrike (pg. 027), and when they did, they “knew with 100% certainty” (pg. 058, n.122) that they were attacking the correct objective and thereby that they were carrying out the intent of the ground force commander. The ground force commander, in turn, believed that “there shouldn’t have been any confusion” about his intent for the requested close air support after he clarified his intent to the JTAC. The JTAC unwittingly created a linguistic bridge between the ground force commander’s stated intent to employ close air support in a manner consistent with “self-defense,” and the “offensive” engagement carried out instead by the aircrew. The JTAC did so by stating that the compound being observed by the aircrew was “currently under the control of the TB [Taliban]” and that the people being observed at the compound were thus “hostile.” They were all trying to “do the exact right thing,” but the training and doctrine they were applying in doing so led to a catastrophic outcome.

Even worse than being ineffective, my concern is that the disciplinary measures imposed will ultimately prove to be counterproductive. The narrative reflected in the official investigation upon which accountability measures were based is itself impermissibly focused on the outcome of the airstrike. How is it, to borrow from one such finding of the official investigation, that the personnel “did not attempt to distinguish between combatants and non-combatants” (pg. 095) but a senior commander can nonetheless maintain that the personnel “were trying to get to the right answer here and do the exact right thing”?

It is because the investigation is manifestly erroneous to conclude that the personnel “did not attempt to distinguish between combatants” and civilians. Doing so would violate the law of armed conflict (“LOAC”) distinction rule. This would constitute a serious violation of international law and would qualify as a war crime. So would engaging in a “facially disproportional” attack (pg. 091), though this official finding fails to present a recognizable formulation of the LOAC proportionality rule upon which it is based. These and similar findings of the official investigation constitute a gross misrepresentation of the applicable factual circumstances and the legal and policy standards from which the investigation draws to develop the relevant conclusions.

Even though the personnel involved in the attack were “trying to do the exact right thing,” as General Votel notes when describing the conduct of the personnel involved in the Kunduz airstrike, “unfortunately they came up short.” The fact that they did “come up short” is attributable to inadequate doctrine and training that persists across the U.S. military as an enterprise. The true causal factors have yet to be identified or corrected, though the personnel directly involved in or otherwise responsible for the attack have received, as General Votel observes, “adverse administrative actions [that] can carry severe repercussions on the careers and professional qualification of these individuals.”

The disciplinary measures imposed following the Kunduz airstrike were not effective in the endeavor to prevent similar tragedies from occurring in the future. The reason is that the measures were directed against individual military members who were inadequately trained by the enterprise that called upon them to use force on its behalf, rather than against the enterprise itself. Even worse, the sense that the personnel involved in the attack are punished for the outcome of the attack even though they were “trying to do the exact right thing” contributes to a prevailing sentiment of enmity toward implementation of military justice in armed conflict. This enmity can actually discourage compliance with existing limitations on the use of force by generating a sense of antipathy toward imposition and implementation of legitimate rules.

To this effect, a reflection captured in one of the sworn statements that is included as an exhibit to the Kunduz investigation should be shared and considered. The affiant, whom I assess based on the context of the statement to be the SOTF-A chief of operations, or “CHOPS,” is asked by the interviewer, “What changes should be made to guidance, SOPs [standard operating procedures], unit procedures or training which could have mitigated” the attack on the MSF trauma center (pg. 436). The initial response of the affiant is, “[t]his is not a simple question to answer because the implication is that there was a breakdown or an issue with the guidance, SOPs, unit procedures or training.”

The reflection that follows this initial answer is lengthy, but it is worth including in full here for consideration. The affiant goes on to observe:

“There is a reason that Special Forces elements were asked to conduct this mission. The experience, maturity, professionalism, and ability to operate in the most extreme and austere environments is why they were chosen. There is risk that commanders have to accept when placing American Soldiers in extremis situations, and when [General] Campbell told [redacted, likely from the context a reference to the ground force commander] to take his forces and seize a foothold in a city that had been seized by an estimated [redacted] INS [insurgents], there was risk associated with this extremely dangerous mission asked of these SF [Special Forces] Soldiers. To highlight the risk, from the moment the USSF convoy departed the airfield, they were in a constant firefight for over 2 days. Every time I spoke with [redacted, likely a reference to the ground force commander] (approximately 6 times from 01 OCT – 03 OCT), I heard constant gun fire and explosions in the background” (pg. 436).

The affiant concludes this reflection by observing, “I believe that any changes/restrictions to guidance, SOPs, or unit procedures as a result of the airstrike would hinder the ground forces’ ability to act in their own self-defense.” This sentiment and the rest of the reflection invoke a conceptual challenge involving the endeavor to achieve accountability and implement effective corrective actions following attacks that result in unintended consequences in armed conflict. While the challenge is inherently conceptual, the attack on the MSF trauma center demonstrates the practical implications of the pervasive conceptual challenge. With the question involving the effectiveness of the specific accountability measures that were implemented after the airstrike addressed, the present inquiry turns to consider some concluding reflections involving matters of broader conceptual and practical concern.

Concluding Reflections

The requirement to utilize force, often lethal force, to achieve an identified outcome sets armed conflict apart from a non-conflict setting. For corrective actions and accountability measures following accidents in armed conflict that result in unintended consequences (such as civilian casualties and fratricide) to be meaningful and effective, both the actions and measures must be developed and implemented with a clear understanding of and appreciation for the operational and information environment that existed before and during the accident. The official investigation conducted following the Kunduz airstrike does not present an accurate assessment of either the operational or information environment encountered before and during the airstrike by the personnel involved in the attack.

While the accounting of the factual record appears to be supported by the evidence gathered during the investigation, the findings are not supported by the evidence derived by the investigation team. Similarly, the official findings are based upon articulations of applicable law and policy that do not present a faithful application of the sources from which the formulations are drawn. Both of these factors render the official findings legally deficient. The corrective actions and accountability measures implemented after the airstrike are based on these findings that are legally deficient, and this impugns their reliability as well.

The assessment conducted in the present inquiry of the circumstances that led to the tragic attack on the MSF trauma center and of the corrective actions and accountability measures implemented is not consistent with the prevailing perspectives expressed by senior MSF representatives in the wake of the Kunduz airstrike. I have studied the MSF reaction and narrative extensively during the course of my assessment of both public and official accounts of the attack. While my assessment of the investigation and subsequent corrective actions and accountability measures does not align with prevailing MSF perspectives, I am committed to the pursuit of both meaningful corrective actions and effective accountability measures to ensure that the errors that caused the horrific outcome of this airstrike do not contribute to another in the future.

I conclude in the full assessment of the Kunduz investigation and lessons “learned” therefrom that the attack on the trauma center does not constitute a war crime, and this conclusion is summarized in the Opinio Juris post of this collaboration. While I explain this conclusion in detail in the full assessment of the investigation, I close the present essay with an observation involving the connection between public narratives and the inherently legal characterization of an attack as a war crime.

This closing observation draws on the ongoing work of Professor Shiri Krebs of Deakin Law School. Dr. Krebs has researched and written extensively about the potential unintended consequences of invoking inherently legalistic terminology such as “war crimes” in public narratives involving actions in armed conflict. She concludes her latest article on the topic with the observation that in our current “fake news” era, “where alternative facts are often generated to counter unwelcomed facts and narratives, it is more important than ever to seek new and better ways to produce and introduce information, and prompt reforms based on the lessons learned.”

This and previous similar observations on the topic by Dr. Krebs have been a guiding force for my current project assessing the Kunduz investigation and lessons “learned” therefrom. The idea for the current cross-platform collaboration was inspired in large part by the call to “seek new and better ways to produce and introduce information.” Like Dr. Krebs, I am concerned that the focus on whether the Kunduz airstrike should be characterized as a war crime may be inadvertently diverting attention away from the institutional reforms that must be implemented following the tragic attack.

Although my assessment of whether the Kunduz airstrike constitutes a war crime diverges from the perspectives expressed by MSF as an organization, I am no less committed to achieving the reforms that are necessary to correct the institutional failures that led to the horrific outcome of the attack and that have thus far been an impediment to the effort by the U.S. military to identify and implement the required reforms.

Nothing short of a sweeping change of culture within the U.S. military as an enterprise is needed. Only this can generate the widespread and systemic reforms that must be implemented to first identify, and then correct the pervasive inadequacies of training and knowledge related to use of force law and policy that led to the tragedy in Kunduz and that persist still today. In this endeavor, I hope to be an ally of MSF rather than an adversary—even if our perspectives do not completely align.

Providing an insider’s perspective of the answers to the unresolved questions posed by MSF is an important first step. The surviving victims of the airstrike, the families of those killed, and MSF as an institution deserve informed answers, even if official answers were not offered. Along with the answers provided in this essay, I declare my commitment to advocate for the institutional reforms that must be implemented and, if permitted to do so, to collaborate with the U.S. military to successfully implement the needed reforms. The victims and families of the attack on the MSF trauma center, as well as MSF as an organization, deserve nothing less.

Editors: Maria Trinidad Alonso Quiros; Lukas Roth

---

More articles from the online collaboration: