European Union

Statement in the Subsidiary Body 5

Developments in Science and Technology Related to Disarmament and International Security, including Artificial Intelligence in the Military Domain

Conference on Disarmament

Geneva, 20 May 2025

Mr. Coordinator,

I have the honour to deliver this statement on behalf of the EU and its Member States. 

The candidate countries North Macedonia, Montenegro, Albania, Ukraine, Republic of Moldova and Bosnia and Herzegovina*, the EFTA countries Iceland, and Norway, members of the European Economic Area, as well as Armenia align themselves with this statement.

We welcome this timely thematic discussion in the Conference on Disarmament Subsidiary body 5 on “developments in science and technology related to disarmament and international security, including artificial intelligence (AI) in the military domain”.

You have our full support as coordinator of this subsidiary body. We also thank Chair of the UNDC Working Group 2, Ms. Julia Rodriguez Acosta, for her useful presentation.

The rapid pace of AI development, and its increasing integration into military systems, is transforming the landscape of international peace and security. As we navigate this transformation, our guiding principle must be clear: the application of AI in the military domain must take place in accordance and full compliance with international law - particularly the UN Charter, International Humanitarian Law, and International Human Rights Law. We also reiterate that human judgment and control over the use of force must always be retained. Humans must remain responsible and accountable, also when it comes to AI in the military domain, to ensure that this technology is applied in a responsible manner.

Mr. Coordinator,

AI presents both opportunities and challenges. It can enhance military operations through increased speed, scale, and precision. It can support decision-making through the processing of vast datasets - whether from surveillance systems, drones, or satellite imagery. AI-driven logistics and predictive maintenance can improve efficiency and reduce operational costs. It can provide greater distance military operations and allow for more precision in uncertain environments.

But these advantages also pose challenges. AI accelerates the OODA loop (observe, orient, decide, act). The increase of speed and scale capabilities may give rise to misperceptions due to inconsistencies between military intentions and the analyses produced by AI-driven systems. This acceleration risks pushing decision-making beyond human control. The greater the speed, the greater the chance of miscalculation, misperception, and unintended escalation. Reliance on decision-support systems in situations related to the use of force requires appropriate safeguards and measures to avoid automation bias. Technology-related challenges such as the quality and veracity of data-sets used in programming and training of AI as well as the potential discriminating impact of algorithmic bias on the reliability of AI systems deserve our full attention.

Mr. Coordinator,

We welcome the continued focus of the United Nations on this matter. We commend the Responsible Artificial Intelligence in the Military Domain (REAIM) process, launched by the Netherlands in 2023, continued by the Republic of Korea in 2024, and soon to be hosted by Spain in 2025. The REAIM "Call to Action" in 2023 and the "Blueprint for Action" - in 2024 endorsed by all EU Member States represent a promising, multi-stakeholder, inclusive approach to global governance. We look forward to a report by the Secretary-General mandated by Resolution 79/239 summarizing the views of Member States on this matter and cataloguing existing and emerging normative proposals.

We also recognize other valuable contributions to the debate, including the AI Action Summit and its Military Talks hosted by France, the “Political Declaration on the Responsible Military Use of Artificial Intelligence and Autonomy”.

These initiatives are complementary. Together, they are helping to build the global momentum we need. The way forward must be collective. We call for deepened international cooperation aimed at understanding the implications of AI in the military domain and identifying viable governance frameworks. We must ensure that multilateral approaches keep pace with technological innovation.

I thank you, Mr. Coordinator.

*North Macedonia, Montenegro, Albania and Bosnia and Herzegovina continue to be part of the Stabilisation and Association Process.