Stop Killer Robots UK is at the informal exchanges on artificial intelligence in the military domain meetings in Geneva between 15 – 17 June 2026. Below are highlights from the second day.
Killer Robots Global Side Event: Autonomy and AI in the use of force: the need for a comprehensive response
Panellists –
- Georgina Hinds, Legal Advisor, International Committee of the Red Cross
- Dr Jessica Dorsey, Co-Director, Realities of Algorithmic Warfare Research Platform, Utrecht University School of Law
- Elizabeth Minor, Head of Policy, Stop Killer Robots
Overview
This side event, organized by the Stop Killer Robots Global campaign, examined current international efforts to regulate Autonomous Weapons Systems (AWS) and broader military applications of artificial intelligence (AI). Speakers reflected on progress achieved within multilateral processes, particularly under the framework of The Seventh Review Conference of the Convention on Certain Conventional Weapons (CCW) in November 2026, the conclusion of the mandate of the UN Group of Governmental Experts (GGE) in December 2026, and the anticipated launch of treaty negotiations in 2027. The speakers felt that these represent the most significant opportunity in a decade to establish international rules governing autonomous weapons, while highlighting significant legal, ethical, humanitarian, and operational challenges that remain unresolved.
The discussion focused on the implications of AI-enabled warfare for civilians, accountability, meaningful human control, and the future development of international law. The speakers emphasized the need for realistic assessments of both the opportunities and risks associated with military AI, grounded in evidence from contemporary conflicts.
Key Themes and Discussions
Progress in International Processes
All three Speakers acknowledged that meaningful progress has been achieved through the CCW and GGE process over the past decade. The discussions have helped translate highly technical debates surrounding autonomous weapons into legal and political concepts that are increasingly understood and accepted by states.
Georgia Hinds noted that the GGE has developed substantial convergent language on autonomous weapons systems, including proposals for restrictions on certain categories of systems and operational contexts. While recognizing these achievements as significant, she cautioned that existing discussions cannot fully address all emerging military AI challenges.
Participants emphasized the importance of taking stock of what current multilateral processes can realistically deliver. While the GGE has advanced discussions on weapons systems, important gaps remain regarding AI-enabled defence support systems and other military AI applications that may indirectly influence the use of force.
The Challenge of Distinguishing Autonomous Weapons from AI Support Systems
A recurring theme was the increasingly blurred distinction between autonomous weapons systems and AI-enabled decision-support tools.
Georgia Hinds highlighted that while states often draw a clear regulatory distinction between weapons systems and defence support systems, operational realities make such distinctions difficult to maintain. AI systems used for intelligence analysis, targeting support, or operational planning can significantly shape decisions regarding the use of force, even if they do not directly control weapons.
Speakers stressed concerns regarding the reliability of AI systems across varying contexts. AI systems remain limited in their ability to understand context, and many suffer from opacity or “black box” characteristics that make their reasoning difficult to explain or evaluate. Participants noted that simply placing a human “in the loop” does not necessarily resolve these challenges if human operators become overly reliant on automated outputs or merely endorse machine-generated recommendations without meaningful scrutiny.
The concept of meaningful human control emerged as a central principle. Participants argued that human control must extend beyond nominal oversight and must involve genuine human judgment over the effects of weapons systems and military decisions.
AI, Remote Warfare, and Civilian Harm
Jessica Dorsey examined the relationship between AI and contemporary forms of remote warfare.
She challenged narratives portraying technologically enabled warfare as precise, clean, or low-risk, arguing that such characterizations often obscure the realities experienced by affected populations. Evidence from recent conflicts demonstrates that civilian harm persists despite technological advances and claims of improved precision.
Participants discussed how AI-enabled targeting systems are often promoted on the assumption that larger datasets will generate better targeting outcomes. However, speakers questioned whether increased data necessarily translates into greater accuracy or compliance with international humanitarian law.
Concerns were raised regarding how legal concepts such as proportionality and excessive civilian harm can be meaningfully assessed when military operations are conducted at increased speed and scale. Speakers warned that the growing mediation of warfare through screens and digital interfaces risks creating both physical and psychological distance between decision-makers and the consequences of military action.
The discussion also addressed the expansion of surveillance practices and behavioural analysis technologies. Participants expressed concern that AI systems may draw conclusions about individuals based on indirect indicators, such as communications patterns or digital associations, potentially reinforcing bias, discrimination, and erroneous targeting decisions.
Speed, Scale, and Accountability
A major concern throughout the event was the military pursuit of speed and scale through AI-enabled systems.
Jessica Dorsey questioned whether the efficiencies promised by AI are being used to improve deliberation and oversight or simply to increase the pace and volume of military operations. Participants warned that accelerating decision-making processes can undermine the careful legal and ethical assessments required under international humanitarian law.
Speakers noted that while AI may promise operational advantages, warfare fundamentally remains a human activity involving death, injury, and destruction. They cautioned that technologies marketed as reducing risk may instead transfer risks onto civilian populations.
The discussion also highlighted the growing role of private technology companies in developing military AI capabilities. Participants raised concerns regarding transparency, accountability, and the extent to which critical decisions in targeting and military operations may be delegated to proprietary systems developed by actors that are not subject to the same legal obligations as states.
Risks of Escalation, Dehumanization, and Loss of Human Judgment
Elizabeth Minor of the Stop Killer Robots campaign emphasized that autonomous weapons and military AI systems create risks of civilian harm, violations of international law, and conflict escalation.
She argued that the increasing pursuit of speed and scale in warfare is inherently dehumanizing and may undermine the moral deliberation that has traditionally accompanied decisions to use force. Participants noted that historical experiences demonstrate how processes of dehumanization can increase the likelihood of atrocities, and warned that AI-enabled systems may amplify these risks.
Minor also drew attention to the growing phenomenon of digital dehumanisation in the use of AI and automated systems in military and security contexts. Dehumanisation can be understood as the denial of individuals’ human qualities, dignity, and individuality, as well as the reduction of meaningful human involvement in decision-making processes. Digital dehumanisation occurs when people are reduced to data points that are then used to inform decisions or actions that affect their lives. She noted that this can result in automated harm—the practical harms arising from digitally mediated decision-making—including violations of privacy, discriminatory profiling, wrongful arrest, arbitrary detention, injury, and death. In situations of armed conflict, where AI-enabled systems are increasingly being adopted by both state and non-state actors, these risks become particularly acute. Minor cautioned that reducing individuals to datasets risks undermining human dignity and moral judgment in decisions involving the use of force, further reinforcing concerns about accountability, civilian protection, and compliance with international law
Speakers expressed concern that military AI systems can operate at speeds that exceed meaningful human cognition and oversight. Risks identified included mass surveillance, discrimination, biased datasets, outdated information, unreliable outputs, and a lack of transparency regarding how systems function and influence decisions.
The absence of transparency regarding the use of AI in military operations was identified as a significant challenge for accountability and independent monitoring. Participants noted that external observers often lack sufficient information to determine whether AI systems were involved in particular military actions.
Future Legal and Policy Developments
The event highlighted growing international momentum for the development of new legal instruments addressing autonomous weapons systems.
Participants noted that more than seventy states have expressed support for advancing discussions toward legally binding measures. Speakers emphasized the importance of prohibiting certain autonomous weapons systems, particularly anti-personnel systems that operate without meaningful human control.
While recognizing that not all militarily significant states currently support new legal instruments, speakers argued that broader international efforts should continue. They stressed that ongoing discussions across multiple international forums indicate increasing recognition of the need for stronger governance frameworks for military AI.
Several participants also called for enhanced transparency, traceability, accountability, and reporting measures to strengthen compliance with existing international law while future regulatory frameworks are negotiated.
Conclusions
The side event underscored both the progress achieved and the significant challenges that remain in addressing autonomous weapons systems and military AI.
Participants welcomed advances made through the CCW GGE process and acknowledged the emergence of increasingly sophisticated legal and political understandings of autonomous weapons. At the same time, speakers emphasized that current discussions do not fully address the broader risks posed by military AI applications, including AI-enabled decision-support systems, surveillance technologies, and targeting processes.
A recurring message throughout the event was the necessity of maintaining meaningful human control over the use of force and ensuring that technological innovation does not undermine human judgment, accountability, or compliance with international law.
Speakers called upon states to continue pursuing stronger international rules and safeguards, including legally binding measures, while ensuring that transparency, accountability, and humanitarian considerations remain at the centre of discussions on the future of warfare.
The event concluded with a call for governments participating in CCW discussions and other multilateral forums to take progressive and concrete steps toward strengthening international governance of autonomous weapons systems and military AI.