How artificial intelligence can support peace

The document argues that AI can strengthen the information basis for peace-building by enhancing conflict data collection, improving the accuracy of conflict forecasting, and supporting scenario modelling for anticipatory action. While acknowledging risks such as surveillance and disinformation, the author suggests that AI tools—including large language models and satellite imagery analysis—can help humanitarian actors and governments better anticipate and mitigate the impacts of armed conflict.

• AI tools, specifically large language models (LLMs), can increase the volume and accuracy of conflict data by processing vast amounts of text, provided they are guided by human experts who understand political complexities.
• Convolutional neural net models can analyze satellite and radar imagery to identify patterns of destruction, such as fires and demolitions, often before news reports can document them, especially in repressed areas.
• Machine learning and statistical modelling can address data gaps by providing probabilistic representations of event timing and locations, and by estimating fatalities in cases where reported numbers are unclear.
• While forecasting the start of entirely new armed conflicts remains extremely difficult because war often results from unobserved bargaining failures, AI is increasingly effective at predicting the trajectories of existing conflicts.
• Modern conflict forecasting is shifting toward capturing full probability distributions to account for uncertainty, allowing models to signal a low but significant risk of extreme escalation even when peace is the most likely outcome.
• AI-driven scenario modelling and simulations can help decision-makers estimate the benefits of conflict prevention and the effectiveness of peacekeeping operations in containing violence.
• AI models can be integrated into 'parametric insurance' mechanisms to trigger immediate humanitarian funding after a destructive event, a method currently being developed by the UN Food and Agriculture Organization (FAO).
• The use of AI for peace is countered by risks that the technology may be used to repress dissent, spread disinformation, and weaken democratic institutions.

APA

Hegre, H. (n.d.). How artificial intelligence can support peace. Peace Research Institute Oslo. https://cdn.cloud.prio.org/files/5d4e83c2-8bf2-46a9-af8f-c53166a39821/How artificial intelligence can support peace PRIO Paper 2026.pdf?inline=true

Chicago

Hegre, Håvard. How artificial intelligence can support peace. Peace Research Institute Oslo, n.d. https://cdn.cloud.prio.org/files/5d4e83c2-8bf2-46a9-af8f-c53166a39821/How artificial intelligence can support peace PRIO Paper 2026.pdf?inline=true.

Wikipedia

{{cite report |last1=Hegre |first1=Håvard |title=How artificial intelligence can support peace |publisher=Peace Research Institute Oslo |url=https://cdn.cloud.prio.org/files/5d4e83c2-8bf2-46a9-af8f-c53166a39821/How artificial intelligence can support peace PRIO Paper 2026.pdf?inline=true |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{hegrendhow, author = {Hegre, Håvard}, title = {{How artificial intelligence can support peace}}, institution = {Peace Research Institute Oslo}, url = {https://cdn.cloud.prio.org/files/5d4e83c2-8bf2-46a9-af8f-c53166a39821/How artificial intelligence can support peace PRIO Paper 2026.pdf?inline=true}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }