UNIDIR today announced the launch of its Centre of Excellence on AI, Peace and Security, a new platform dedicated to strengthening global governance of artificial intelligence in the context of international peace and security.
As AI transforms security dynamics, from military operations and critical infrastructure to the information environment, governments and institutions are grappling with how to govern its risks and opportunities. The new centre responds to this challenge by bringing together analysis, dialogue and practical support under one roof, serving as the umbrella for UNIDIR’s work on AI and security.
To learn more, we spoke with Dr Robin Geiss, UNIDIR’s Director, and Dr Giacomo Persi Paoli, UNIDIR’s Head of Security and Technology Programme.
Q: AI is advancing at extraordinary speed. From a peace and security perspective, what developments concern you most, and what new risks are they creating for international security?
RG: The pace of advances in artificial intelligence is remarkable. While it brings significant benefits, it is also creating new challenges for international peace and security.
The concern is less about any single application, but rather the cumulative impact across multiple domains.
AI is now being integrated into military systems, cyber operations, critical infrastructure, and the information environment. This can enhance efficiency and decision-making, but it can also increase complexity, create new risks and pathways for escalation, increase the pace of battle, and add new layers of uncertainty in a highly dynamic global security environment. AI can lower the barriers to sophisticated cyberattacks, enable more convincing disinformation, and heighten the risk of miscalculation or unintended escalation in crisis situations.
At the same time, governance and regulatory frameworks are struggling to keep pace. This creates a growing gap between technological capabilities and our collective ability to understand and manage their implications.
Addressing these challenges requires more than technical solutions. It calls for sustained dialogue, shared understanding, and practical cooperation among governments, industry, researchers, and civil society to ensure AI strengthens – rather than undermines – international peace and security.
Q: Governance of AI is often described as fragmented. Where do you see the most pressing challenges today, and why is addressing them becoming increasingly urgent?
RG: The challenge is not a lack of effort. Across the world, a wide range of stakeholders are actively working to shape AI governance. The problem is that these efforts often unfold in parallel – across different communities, processes and policy domains that do not connect effectively.
AI issues are being discussed across multiple domains, from international security and arms control to human rights, development and digital governance. While each track is important, this dispersion creates a fragmented landscape, making it harder to maintain a shared understanding of risks, opportunities and emerging policy approaches.
This fragmentation is becoming increasingly urgent as AI is already embedded in systems that affect security, stability and societal resilience. Policymakers face growing pressure to act in areas where technology, policy and security intersect – often without a fully aligned picture.
The goal is not a single, unified framework. It is better coordination, stronger capacity, and more informed dialogue so that the responses are coherent, inclusive and effective.
Q: Why is UNIDIR launching a Centre of Excellence on AI, Peace and Security now?
RG: UNIDIR is launching its Centre of Excellence on AI, Peace and Security at a pivotal moment, as artificial intelligence becomes central to international peace and security. In the coming weeks alone, Geneva will host two major gatherings – one focused on military application of AI and another one on global AI governance – bringing together Member States and a broader community of actors. The centre therefore responds to a clear need for sustained analysis, inclusive dialogue, and practical support.
While important governance initiatives are emerging around the world, many stakeholders still face capacity constraints, and there remains a gap between policy discussions, technical expertise, and real-world implementation.
This launch also reflects UNIDIR’s own evolution. In recent years, the Institute has become a trusted platform for research, dialogue and capacity-building on AI and security. The centre consolidates and scales this work, providing a more focused and sustained contribution to global efforts on AI, peace and security.
Q: There is no shortage of AI initiatives and forums. What makes this centre different, and what unique value can it bring?
GPP: It is true that there is no shortage of initiatives focused on AI, but relatively few focus on the most consequential and sensitive dimensions: international peace and security.
Many existing platforms centre on innovation, economic growth, digital governance, human rights, and broader societal impacts. These are essential conversations, but issues such as military applications, strategic stability, conflict dynamics, and the future of warfare have often been addressed in more specialized – and fragmented – settings.
The centre is designed to close that gap. It focuses on the intersection of AI, peace and security, examining risks and opportunities associated with AI-enabled capabilities, crisis decision-making, and international stability.
At the same time, many of the most advanced practices on AI governance – such as safety, accountability, testing, risk management, and responsible use – are emerging outside the security domain. A core objective of the centre is to bridge these worlds, connecting expertise, and transferring insights across communities that rarely engage directly.
In doing so, it seeks to strengthen understanding, reduce fragmentation, and support more coherent approaches to AI governance in the security domain.
Q: UNIDIR’s Centre of Excellence for AI, Peace and Security will serve as the umbrella for the Institute’s work on AI. What are its initial priorities, and where do you expect the centre’s work to have the greatest impact in the years ahead?
GPP: The centre builds on several years of UNIDIR’s work at the intersection of AI and international security – bringing research, dialogue and capacity-building under a single, dedicated framework. This allows the Institute to scale its contribution at a time of rapidly growing demand and expertise.
Its initial priorities span three areas: deepening analysis, strengthening engagement, and delivering practical tools. On the research side, this includes emerging issues such as agentic AI, cross-domain security implications – from cyber to biosecurity – and the risks linked to proliferation.
Engagement will also expand, especially with industry, recognizing the critical role that developers and deployers play in shaping security outcomes. In parallel, the centre will produce practical outputs, such as guidance for national AI and defence strategies, and an AI security lexicon to support clearer, shared understanding across communities.
Looking ahead, its greatest impact will come from its ability to act as a bridge: between policy and technology, across security and non-security communities, and from global debates to real-world implementation.
As AI reshapes security dynamics, the need for trusted platforms that enable informed, inclusive and actionable policymaking will only grow. That is the space the centre aims to occupy.
Q: The centre is built around three pillars: knowledge, dialogue and action. How will these translate into practical support for governments and other stakeholders?
GPP: The three pillars reflect our belief that effective AI governance in the peace and security domain requires more than analysis alone; it requires connection and implementation.
- The knowledge pillar focuses on generating policy-relevant research on emerging issues, helping decisionmakers understand risks, opportunities and trends.
- The dialogue pillar creates trusted spaces for exchange among governments, industry, academia, civil society, and international organizations – bringing together communities that do not often engage and fostering shared understanding and trust.
- The action pillar turns insights into practice. It focuses on tools, guidance, training and capacity-building – helping translate principles into concrete policies and operational measures.
Together, these pillars are designed to equip governments and other actors not only to understand AI’s implications for peace and security, but to respond collaboratively, coherently and effectively.
Q: Geneva has long been a hub for diplomacy and international cooperation. What makes it particularly well suited to host a global centre focused on AI, peace and security?
RG: Geneva is uniquely placed at the crossroads of international security, technology governance, humanitarian action, and human rights – fields that are increasingly intertwined in the age of AI. Few places in the world bring together such a diverse and influential mix of actors working across these domains.
The city hosts international organizations, diplomatic missions, technical experts, humanitarian actors, standards-setting bodies, and civil society organizations that are shaping discussions on the future of technology and its societal impacts. It has also emerged as an important centre for AI governance, with a growing ecosystem focused on digital technologies, responsible innovation, and international cooperation.
For a centre focused on AI and security, this convergence is a decisive advantage. The implications of AI extend far beyond traditional security institutions, touching everything from human rights and development to cyber resilience and humanitarian action. Geneva provides a unique environment in which these different perspectives can be brought together.
By anchoring the centre here while engaging globally, UNIDIR can leverage this ecosystem to connect communities, deepen dialogue, and support more coherent and inclusive approaches to AI governance in the peace and security domain.
