The Data for Peace Conference 2026 (June 15–17, Stockholm and online) brought together the research, technology, policy, and practitioner communities working on conflict prevention. Two in-person plenary days in Stockholm—attended by 150 participants on-site—were followed by a virtual day of parallel tracks that drew more than 1,500 registered participants. Session recordings from the three days have since generated over 1,800 unique views, and counting. Formats ranged from panels and lightning talks to workshops, peer-review clinics, and scenario simulations—designed not only to showcase what is technically possible but also to ask what is required for innovations to be applied responsibly and effectively, turning warnings into action and action into peace.
The Conference was co-hosted by the Department of Peace and Conflict Research at Uppsala University, the NYU Center on International Cooperation (CIC), and AI for Peace—with support from the Complex Risk Analytics Fund (CRAF’d), Folke Bernadotte Academy (FBA), and the Finnish Ministry for Foreign Affairs.
The Conference generated the richest body of content the Data for Peace initiative has produced to date, and no summary can do justice to the full breadth of what was shared across both in-person and virtual days. This is not a comprehensive account but a thematic synthesis—an attempt to trace the threads that connected sessions across three days, rather than a reflection of any single session’s full content. We begin by looking at where the field stands on peace and conflict data—how and whether AI is helping, where conflict modeling is headed, and whether the field is finally closing the gap between warning and action. We then turn to the harder questions: what trust, accountability, and political will are required to make warnings actionable—and who early warning systems ultimately serve and belong to.
For the full picture, we invite you to explore recordings of individual sessions on our YouTube channel or of the complete conference days (Day 1 and Day 2).
