International AI Governance

Artificial Intelligence (AI) is an emerging technology which presents unprecedented risks and opportunities. In the international security domain, rapidly advancing AI capabilities could cause catastrophic outcomes via several pathways, including misuse by malicious actors, accelerating arms races and the erosion of human control. Yet, with appropriate international coordination and safeguards, advanced AI could contribute immensely to scientific progress, crisis management and global peace.

This working group explores three key areas at the intersection of nuclear disarmament and AI governance:

AI and international security: Studying how advanced AI affects strategic stability and the risk of nuclear conflict, for example via autonomous weapon systems, shifts in offense/defense balance, AI integration in military decision-making, or competitive dynamics in AI development. Also, how geopolitical dynamics and security concerns impact the prospect of international cooperation on AI safety and governance.
Lessons from nuclear governance: Drawing on the history, theory and mechanisms of nuclear governance —e.g., arms control, international institutions, the logic of deterrence—to inform AI governance. This also includes methodological contributions, e.g., towards a more systematic approach to learning from analogies.
AI in nuclear disarmament: Exploring opportunities such as leveraging advanced AI to enhance verification regimes, improve transparency and communication, and support cooperation between states. Additionally, considering the risks from the application of advanced AI in this high-stakes context.

Working group Leader: Sophia Hatz

Sophia Hatz is an Associate Professor with 11 years of experience in peace and conflict research. Her current research interests include nuclear disarmament, international governance, coercion and deception, and risks from advanced Artificial Intelligence (AI). Hatz has extensive experience with data collection, research design, causal inference and quantitative analysis of large-n datasets.

Publications

2025

Adrien Abecassis ... Sophia Hatz, et al. 2025. A Blueprint for Multinational Advanced AI Development. Oxford Martin AI Governance Initiative.
Hatz, Sophia. In press. The Nuclear Analogy in AI Governance Research. In Handbook on Global AI Governance, edited by Magnus Lundgren and Markus Furendal, Edward Elgar Publishing. [Chapter accepted by editors; final handbook pending publisher acceptance]
Hatz, Sophia, Noemi Dreksler, Baobao Zhang, and Kevin Wei. 2025. “Local US officials’ views on the impacts and governance of AI: Evidence from 2022 and 2023 survey waves.” PLoS One 20(10): e0332919. https://doi.org/10.1371/journal.pone.0332919

Contact

For questions about the working group's research and activities:
Sophia Hatz