The US has several relevant policy documents that comprehensively outline its CHMR policy and approach. Most notable are the 2022-2025 CHMR-AP and the 2023 DoD-I, which contain detailed descriptions on US policy on preventing, tracking, investigating, and responding to harm. These build on a series of earlier publications reviewing US efforts to mitigate harm;  however, the CHMR-AP and DoD-I are distinct from past policies and documents in placing CHMR at the heart of strategic and operational thinking, their ambition in reforming the enterprise-wide approach to CHMR, and their willingness to acknowledge and address existing gaps. ¹⁷

The CHMR-AP, DoD-I, and earlier documents were built on significant input from civil society and independent experts, through both formal and informal forums. ¹⁸ While not all recommendations were brought into the final policies, this process ensured that the DoD had access to independent expertise and more comprehensive lessons on the civilian harm from its actions. 

In addition to the CHMR-AP and DoD-I, the 2018 National Defence Authorization Act Section 1057 requires the DoD to report annually to Congress on civilian harm from its military action. These reports must identify operations in which civilian harm was confirmed or reasonably suspected; specify dates, locations, theatres of conflict, and operation types; assess civilian and combatant casualties; and describe investigative processes and mitigation measures. Subsequent NDAA amendments in the 2019 NDAA expanded reporting requirements to include mission-level detail and distinctions between civilians killed and injured. ¹⁹

The current challenge in the US context centers not on the policies that exist, but on the apparent lack of appetite, willingness, and capacity to implement these policies in the context of significant budget and staff reductions and a focus among DoD leadership on ‘lethality’ over legality or legitimacy. As an example of this, while the CHMR-AP officially ‘ended’in September 2025, many of the important improvements that were emerging in the US approach as a result have since been curtailed in the midst of budget-cuts, with damning consequences for those caught in armed conflict, from the Caribbean to Yemen. ²⁰

Beyond challenges to DoD policies on CHMR from the current Trump Administration, it is also notable that the US has never had a sufficient policy framework on civilian harm mitigation or response from non-military actors like the CIA who regularly use deadly force overseas. 

US policy documents contain clearly worded commitments to protecting civilians, as well as clear definitions and explanations of key terminology. Both the CHMR-AP and the DoD-I 3000.17 contain comprehensive and tangible language, signaling strong policy intent. The DoD has also set up an accessible central webpage with an overview of all CHMR-related policies, adding useful clarity on what is considered part of the US policy framework on CHMR.  

The DoD-I includes a comprehensive definition of ‘civilian harm’ (textbox 2). ²¹ The existence of such a definition, and its comprehensive inclusion of both direct and indirect harms, constitute best practice internationally. 

The CHMR-AP and DoD-I expressly identify specific roles and responsibilities for different entities within the DoD, often in great detail; in the years leading up to the second Trump Administration, many of these roles were established, including through the Civilian Protection Center of Excellence (CP COE), as well as Civilian Harm Assessment Cells, and Civilian Harm Mitigation and Response Officers (CHMROs) throughout the US Unified Combatant Commands (CCMDs). ²²

The CHMR-AP delineates the timeline of this integration by fiscal year to serve as guideposts for implementation, enhancing transparency and the potential for public oversight and monitoring. The CHMR-AP also includes specific objectives for integrating CHMR into joint doctrine and operational planning, as well as objectives to train personnel in CHMR and embed CHMR Officers across the DoD.

However, the resourcing of the roles established by the CHMR-AP and DoD-I is dependent upon Congressional appropriations through the annual National Defence Authorisation Act (NDAA) and other budgetary tools, which means that the implementation of CHMR policy and practice are subject to evolving political interests (see indicator 1.4). 

In 2025, the CHMR enterprise as a whole across the DoD and the operational CCMDs saw a significant reduction in staff. This affected CHMR-specific functions, such as the Civilian Protection Center of Excellence, as well as other functions within the DoD that are vital for appropriate civilian harm mitigation, including legal advisors.

Over the last year, there have been drastic cuts to the teams and staff responsible for effective CHMR in the US setting. Staff who were critical to the implementation of US CHMR systems have been removed from their roles or re-assigned to non-CHMR roles. This puts the DoD’s ability to effectively mitigate harm to civilians – and respond when it occurs – at risk. 

Even as staff were removed from their positions, the budget available remained relatively steady; as a result, the budget appears relatively strong on paper, but there is no capacity to implement projects under the relevant budget line.  The DoD has received specific funding from Congress for CHMR activities and personnel in the past. This included $41.75 million in fiscal year 2023 and $41.5 million in fiscal year 2024 to the DoD for CHMR activities and personnel. ²³ In 2025, the US government was funded through a continuing resolution, so funding levels from 2024 were repeated for 2025. ²⁴

Congress has further set up an annual fund from which ex gratia payments for US-caused civilian harm can be paid. This fund, which was established in 2020, contains an annual $3 million fund for the provision of ex gratia payments in response to civilian harm caused by US military operations. ²⁵ In 2023, the DoD submitted to Congress a legislative proposal titled “Payment of Expenses to Help Respond to Civilian Harm.” The proposal creates an additional $10 million authority for DoD to respond to civilian harm through a variety of in-kind assistance options. This authority would supplement, not replace, the existing $3 million ex gratia authority. ²⁶ However, this proposal was not passed into either the NDAA 2024, nor the NDAA 2025 according to summaries and legislative text reviews. ²⁷

CHMR-AP Objective 11 focuses on the staffing needed to implement the action plan objectives, including a manpower study to ensure the projected staffing needs are accurate. It also includes annual funding requests to Congress and periodic updating of anticipated resource requirements. The findings from the initial study, which are not public, were presented to the CHMR Steering Committee in July 2023; one purpose of the manpower study was to validate the assessment in the CHMR-AP that 166 FTEs would be needed to fully implement the action plan. ²⁸ However, in the time since the manpower study, the Trump Administration has significantly reduced the staff focused on CHMR.  ²⁹

In spite of evidence of significant levels of harm from US actions, the US has not announced the outcome of any investigations into harm that occurred that year. 

The DoD has had the capacity to investigate allegations of civilian harm in some recent conflicts, which has led to the acknowledgement of civilian deaths, injuries to civilians, and damage to civilian infrastructure. Several CCMDs have in place reporting mechanisms and tracking practices, which feed into investigations. However, there is no evidence available in the public sphere to suggest that these were being applied in US operations throughout 2025, particularly those targeting small boats in the Pacific and Caribbean, as well as significant allegations of harm in Yemen in 2025, which have not yet been publicly reported on by the DoD. 

Beyond these contexts, there is often a significant discrepancy between the evidence on civilian harm from open sources and the level of harm recognised by the US. The US-led counter-ISIS coalition, for instance, has killed at least 8,262 civilians according to Airwars’ documentations; the coalition had acknowledged 1,457. ⁴¹

The DoD reports to Congress annually on its investigations into civilian harm allegations (both those which have been deemed “reasonably suspected” of being caused by the US, and those rejected by the DoD). Congress mandates this reporting in Section 1057 of the FY2018 National Defense Authorization Act; the DoD has released these reports, often referred to simply as ‘1057’ reports, annually since 2022. While often significantly delayed up to a year, each of these reports has offered important insight into the DoD approach to investigating allegations of harm, the capacity that exists to conduct such assessments, and a brief overview of the reasons why some allegations have been rejected. Beyond its acknowledgement or rejection of civilian harm, the 1057 reports also include an overview of how many allegations have been brought to the attention of the DoD. They are limited to reporting on allegations of harm from actions where a US service member deployed a weapon (and therefore does not account for the often significant harm from US support for allies and partners) and it does not cover covert operations such as those conducted by the CIA in Yemen. 

Over the last four years, the US has made efforts to further codify its approach to civilian harm investigations, particularly through the CHMR-AP (2022), the DoD-I 3000.17 (2023). DoD-I 3000.17 provides procedures for conducting civilian harm assessments. In addition, in preparation for and throughout the duration of crisis or conflict, it requires combatant commands to establish or maintain civilian harm assessment cells (CHACs) and to have a Civilian Harm Assessment and Investigation Coordinator in place at the combatant command-level, or, as appropriate, at subordinate operational commands. However, in light of budgetary cuts and reductions in CHMR-related staff across the DoD, the practical implementation of these positions has been severely undermined. ⁴²

In the last few years, the US DoD has shifted from using language of ‘credible’ or ‘non-credible’ on its designation of civilian harm allegations, to ‘more likely than not’ threshold. This is a positive step, offering a more appropriate threshold of evidence for warzones. Yet the path towards reaching the conclusion of ‘more credible than not’ does not appear to follow a transparent set of criteria that is used across CCMDs. 

In cases where the US determines that it has not caused civilian harm because it believes those targeted to be combatants, it offers little explanation for its determinations, even when there is repeated testimonies that those killed were civilians. This is particularly true when it comes to the military operations targeting small boats in the Caribbean and Pacific in 2025, where the US Administration sits at odds with legal experts and international bodies (including the UN) as to the legal frameworks and justification for the strikes. ⁴³

The DoD-I 3000.17 (2023) and AR 15-6 regulation mention site visits and witness interviews as possible investigative means. However, in practice and particularly in air campaigns, site visits have been extremely rare. Policy further does not specify that witness interviews should include civilians. In the air campaigns that have defined US warfare for over a decade, the US has not often utilised investigative methods like site visits and witness interviews, even though in some cases there were Special Operations Forces (SOF) personnel present on the ground, e.g., in Somalia. The Center for Civilians in Conflict and Columbia Law School Human Rights Institute found that between 2016 and 2019 not a single civilian witness was interviewed or a single site visit conducted in any investigation of civilian casualty allegations in Somalia. ⁴⁴

There is no publicly available evidence to suggest site visits have been conducted by the US DoD in 2025. In past 1057 Reports, the DoD has specifically clarified that it has not conducted site visits or witness interviews. ⁴⁵

The DoD-I sets policy for how civilian harm assessments and investigations should be conducted, including that civilian harm assessment cells (CHACs) should use “all available information,” ²¹ such as interviews with witnesses or open-source reports where feasible. However, the policy does not require public disclosure of when such methods are actually used and there is no requirement for the DoD to publish detailed operational data like specific site visits or civilian interviews in unclassified form. 

The US did not report on investigations into the numerous allegations of harm which emerged in 2025. In July 2025, the 2024 1057 Report was released; it included expected levels of detail, with extensive sections dedicated to assessment and investigative processes, as well as age and gender disaggregated results on civilian casualties. ⁴⁷

The United States has multiple policies and procedures in place to promote consistent public reporting on civilian harm and civilian harm investigations arising from US military operations. A central pillar is Section 1057 of the National Defense Authorization Act (2018), which requires the DoD to publish annual unclassified reports on civilian casualties. ⁴⁸ These reports must identify operations in which civilian harm was confirmed or reasonably suspected, specify dates, locations, theatres of conflict, and operation types, assess civilian and combatant casualties, and describe investigative processes and mitigation measures. Subsequent NDAA amendments expanded reporting requirements to include mission-level detail and distinctions between civilians killed and injured. ⁴⁹

Complementing statutory reporting, DoD-I standardises terminology and public-facing communication. It directs the DoD to avoid credibility binaries and instead assess whether civilian harm was “more likely than not,” refer to “information” rather than “allegations,” and coordinate the release of outcomes through public affairs channels. The instruction also commits the DoD to publishing releasable information (such as date, location, and type of operation) when civilian harm occurs.

At the operational level, Combatant Commands are required by the DoD-I to publish at least quarterly unclassified civilian harm reports. ²¹ There is, however, inconsistent implementation of this and reporting templates are not standardised across commands, a full overview of investigations are not published, methodological reasoning is – at times – opaque, victim notification is not required, and the role of intelligence agencies in strikes remains unreported. ⁵¹ 1057 reports have repeatedly been significantly delayed, delaying accountability and transparency. 

In practice, the connection between investigations and concrete responses to civilian harm is weak in the US. While the DoD routinely acknowledges civilian harm – an important and positive practice – public reports often stop at acknowledgement and do not explain what follow-on actions, remedies, or amends are planned. Commanders appear to lack clear, publicly articulated guidance on the range of available response options, such as ex gratia payments, apologies, or other forms of redress. As further explored in Category 5, while the US has a $3 million annual funding pool available for ex gratia payments, it made only a single payment between 2020-2022. ⁵² There is no public evidence that it has made ex gratia payments in subsequent years. ⁵³

This practice is different from what is set out in policy. Recent US policy documents explicitly link civilian harm investigations to acknowledgements of harm, responses, and operational learning. Notably, the DoD-I states that CCMDs or subordinate commands should, as appropriate, use information from civilian harm assessments and investigations to inform acknowledgements and other responses to civilian harm. The instruction also establishes processes for capturing lessons learned, recommendations, and practices related to mitigating and responding to civilian harm, including those derived from investigations conducted under the CHMR framework.

Although US policy formally provides for criminal investigation, practice reflects limited commitment and transparency. During 2025, leadership in the DoD, particularly Secretary of Defense Secretary Pete Hegseth, signaled a rhetorical and practical distancing from both international and national law. ⁵⁴ This shift was reflected less in formal withdrawal from legal obligations than in an apparent de-prioritisation of international humanitarian law language, expressed skepticism toward external scrutiny, and a stronger emphasis on commander discretion and operational freedom, with little recognition of the military operational value of effective CHMR. While core legal frameworks formally remain in place, statements made throughout 2025, and the removal of key legal staff, suggest a narrowing of how legality and international accountability are understood and operationalised within the DoD. ⁵⁵

This points to yet another significant gap between policy and practice in the US. There is a strong policy-foundation for criminal investigations in allegations of civilian harm which may be of a criminal nature; the CHMR-AP, for instance, mandates that if evidence emerges suggesting civilian harm resulted from criminal activity, civilian harm assessments must be suspended, evidence preserved, and the responsible Military Criminal Investigative Organization (MCIO) immediately notified. Legal and policy frameworks governing criminal investigations are well established, and as noted by the Center for Civilians in Conflict and the Columbia Law School Human Rights Institute, incidents involving potential criminal violations are meant to trigger mandatory MCIO investigations. ⁵⁶

In practice, however, criminal investigations related to civilian harm are exceedingly rare and are seldom reported publicly. When they do occur, cases are typically pursued under ordinary domestic criminal offenses (such as murder or assault) rather than explicitly under the Law of Armed Conflict or international humanitarian law. ⁵⁷ There appears to be no distinct or transparent process for investigating IHL violations as such.

Historically, independent human rights organisations have documented serious allegations involving US forces in Afghanistan and Iraq, including abuse during interrogations (most notably Abu Ghraib prison) and mass civilian casualty incidents such as Haditha massacre. ⁵⁸ These cases resulted in few criminal convictions, and some convictions were later reduced or overturned through clemency. ⁵⁹

Structural accountability gaps also persist. The United States is not a party to the Rome Statute, and domestic legislation such as the American Service-Members’ Protection Act limits cooperation with the International Criminal Court, further constraining avenues for criminal accountability.

Amends processes and procedures for civilian harm claims are embedded throughout a number of policy documents, as well as in legislation – however, the gap between policy and practice is stark. While 611 ex gratia payments were made in 2019, for instance, only 3 payments have been confirmed publicly since 2020; in the same period, the US DoD has caused unprecedented harm in some of its theatres of engagement. ⁷⁵

The lack of monetary responses to civilians affected by warfare is particularly troubling in light of the existence of an annual $3m fund created by the FY2020 National Defense Authorization Act and extended until 2033 by the FY2022 NDAA. ⁷⁶ It is therefore clearly not a lack of means to respond to harm that is preventing a monetary response, but rather insufficient systems and lack of will. 

The US has additional mechanisms in place, such as the Foreign Claims Act (FCA), which does not govern combat operations but does provide recourse for civilians to claim compensation if harmed overseas by US military personnel. ⁷⁷

Prior to the new administration taking office in January 2025, a number of plans were underway to develop additional guidance and framework for amends processes. Due to widespread cuts across the DoD, the team responsible for developing this guidance was significantly reduced. ⁷⁸ Commands that already had a process for receiving requests about civilian harm maintained that capability, though with reduced staffing capacity. 

When SOUTHCOM began its military campaign against drug smuggling routes in the Caribbean and Pacific, a system for soliciting information about civilian harm was not put in place, preventing a response to harm. ⁷⁹ Pre-existing language in both the NDAA and the FCA in any case already included exclusions for amends for those ‘unfriendly’ to the United States, which will serve to further challenge recourse to amends for those harmed in these strikes through DoD channels. Instead, in at least one case the families of individuals killed in the strikes have resorted to suing the US, with representation from the American Civil Liberties Union. ⁸⁰

The DoD submitted a legislative proposal to Congress in 2023 titled “Payment of Expenses to Help Respond to Civilian Harm.” ²⁶ The proposal creates an additional $10 million authority for DoD to respond to civilian harm. This authority would supplement the existing $3 million ex gratia authority. However, this proposal was not passed into either the NDAA 2024, nor the NDAA 2025 according to summaries and legislative text reviews. ²⁷

There were no public acknowledgements in 2025 of harm resulting from the US campaign that year, including in Yemen from March to May; according to open-source evidence, this campaign led to unprecedented levels of civilian harm in the country from US actions. ⁸³

In 2025, the US publicly acknowledged responsibility for harm that has occurred in previous years, including in the annual 1057 report to Congress released in July, and in individual reports released by commands. ⁸⁴ Acknowledgements included references to historic incidents of harm, indicating that work is on-going within commands to review past cases. These acknowledgements reflected requirements set out in the NDAA to release a report to Congress every year, and reflected requirements outlined in DoD-I on routine command level reporting obligations. 

Despite these ad hoc legacy admissions, as outlined in Section 4 of this report, requirements around reporting on civilian harm are subject to the discretion of the Commander, who may choose to delay or withhold reports to protect classified or sensitive information. For the Yemen campaign, the applications of these exemptions with regards to civilian harm incidents have not been publicly stated. 

The CHMR-AP outlines the aim to publish public affairs guidance emphasizing the proactive release of information and tailorable public affairs approaches regarding civilian harm, which demonstrates respect for civilians and communities harmed by US operations and transparency regarding DoD policies and processes for mitigating and responding to civilian harm. ⁸⁵ The status of this policy development is currently unknown. ⁸⁶

There were no public acknowledgements in 2025 of harm resulting from the US campaign in Yemen from March to May; this campaign led to unprecedented levels of civilian harm in the country from US actions. Without acknowledgement, there cannot be a response. 

Response options are listed in the CHMR-AP in broad terms, and acknowledge that there may be specific cultural and contextual considerations that should be taken into account in responding to civilian harm claims. The development of policies governing these options remain unknown, though it is likely that further research into these topics has been challenged by reductions in staffing and resources on this subject following department-wide cuts under the new administration. ⁸⁷ This means that critical gaps identified in the early policies on this subject, such as linkages between strategic effects and condolence payments, have not been addressed. ⁸⁸