[Dr Emily Camins is a lecturer in international humanitarian law at the University of Western Australia Law School.]
On 6 July 2026, a small AI-guided Russian drone crashed into a wall and exploded in the town of Zaporizhzhia, Ukraine, killing 19-year-old university student Tetiana Bubynets, as well as Oleksiy Svirin, 41, and Roman Karpiy, 48. The drone had been programmed by human operators to travel towards a specific gas station, but appears to have selected its precise target (possibly propane tanks) on its own. Subsequent investigation and analysis indicated the drone had been guided by an experimental AI system using a computer chip sold by Nvidia.
Further civilian harm from widespread use of lethal autonomous weapons systems (abbreviated, somewhat problematically, to LAWS) in armed conflict is likely. While some argue that LAWS might lead to greater precision on the battlefield, there is strong opposition to the technology, with experts warning of a risk of additional targeting errors, discrimination, murky accountability frameworks and desensitization to the effects of armed conflict. In any event, even lawfully deployed weapons can cause death and injury to civilians and destruction of civilian objects (‘awful but lawful’ harm). To the extent that LAWS lower the threshold for armed conflict – itself a contested proposition – the risk of civilian harm from such weapons is heightened.
Accordingly, measures that support victim-survivors of LAWS are important. Yet to date, this issue has received limited academic or governmental attention. For example, the Report of the 2024-2025-2026 sessions of the Group of Governmental Experts (GGE) on Emerging Technologies in the Area of Lethal Autonomous Weapons Systems CCW/GGE.1/2026/CRP.1/Rev.1 (03 September 2026) addresses doctrinal questions of accountability for autonomous systems (see eg paras 42-44). Similarly, the 5 June 2026 rolling text of the GGE on LAWS, like its predecessor of 12 May 2025, focuses on ensuring States, parties to armed conflict, and individuals remain ‘responsible … and accountable’ for decisions with regard to LAWS (chapter V). Previous working papers have taken a similar approach. To the extent that victims have been considered in the GGE discussions (such as in the 2024 working paper submitted by Bulgaria, Denmark, France, Germany, Italy, Luxembourg and Norway (CCW/GGE.1/2024/WP.3)), it has been in the context of holding States responsible for internationally wrongful acts or violations of IHL (para 10). Reparation or other measures directly or expressly in support of victims of LAWS have been largely absent from the discussion.
While compensation is often the primary reparative response in cases involving violations of IHL, it is by no means the only one, nor necessarily the most important. The Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA) set out three forms of reparations: restitution, compensation and satisfaction. Restitution requires a State to ‘re-establish the situation which existed before the wrongful act was committed’ (Article 35). Restitution is often unavailable as a response to violations of IHL, particularly those involving death or personal injury. The compensation obligation (which is reflected in Article 91 of Additional Protocol I and customary IHL) requires States to compensate for any financially assessable damage caused by an internationally wrongful act (Article 36). The satisfaction obligation may involve an acknowledgement of the violation, a statement of regret, an apology or another appropriate measure (Article 37). To this somewhat State-centric list, the victim-survivor-oriented Basic Principles and Guidelines on the Right to a Remedy and Reparation for Victims of Gross Violations of International Human Rights Law and Serious Violations of International Humanitarian Law add rehabilitation and guarantees of non-repetition. These important measures are, like reparation under the ARSIWA, based on the existence of violations of IHL and international human rights law.
Compensation is directly referable to questions of accountability and responsibility for LAWS, which emerging LAWS frameworks are already seeking to address. Despite these efforts, the use of LAWS is likely to make the path to reparations more challenging by obscuring who made the decision that caused harm. Even where responsibility can be established, questions beyond legal liability also arise. For example, where LAWS make erroneous targeting decisions based on opaque data or reasoning, how can a State meaningfully guarantee non-repetition? Likewise, where civilian harm results from opaque LAWS decision-making, how meaningful can an apology be? These questions are only partially addressed by the GGE Report, which emphasizes the importance of human control and judgment.
More generally, States are sometimes reluctant to confer reparative rights directly on individuals, but have proven more willing to create alternative victim-centric obligations. As a complement to general obligations relating to State responsibility, humanitarian disarmament treaties such as the 2008 Convention on Cluster Munitions (Article 5) and the 2017 Treaty on the Prohibition of Nuclear Weapons (Article 6) have adopted victim assistance frameworks. In contrast with rules of State responsibility, these treaties impose primary obligations on States to provide assistance, including medical care, rehabilitation and psychological support, as well as economic and social inclusion measures, to victims of the eponymous weapons within their jurisdiction or control. This obligation extends to family members of those killed and others. As primary obligations, these provisions obviate the need for victim-survivors to establish legal liability for violations of the law. In the context of LAWS, where responsibility may be more contestable, this approach would offer victim-survivors more accessible support than measures based strictly on principles of State responsibility.
There is precedent for adopting a victim assistance approach within the Convention on Certain Conventional Weapons framework. Building on the 1997 Anti-Personnel Mine Ban Convention (Article 6(3)), Protocol V on Explosive Remnants of War (2003) provides in Article 8(2) that States parties ‘in a position to do so shall provide assistance for the care and rehabilitation and social and economic reintegration of victims of explosive remnants of war’. Introducing a victim assistance provision into the GGE process would be a modest measure that would help to ensure that those who are likely to be most harmed by LAWS are not forgotten as autonomous weaponry reshapes warfare.
