L'intelligence artificielle au service de la consolidation de la paix

The document examines the dual-use nature of AI in peacebuilding, highlighting its potential to improve early warning systems, civic engagement, and humanitarian mapping while warning against risks such as digital repression, algorithmic bias, and the erosion of human oversight. It argues that for AI to support peace rather than instability, it must be governed by human rights principles, transparency, and community participation.

• AI-driven early warning systems, such as Violence in Early Warning Systems (ViEWS) and the Armed Conflict Location & Event Data (ACLED) project, can predict conflict with short-term accuracy up to 95% by identifying indicators like hate speech trends and troop movements. However, these models are prone to errors and biases in data-poor environments, meaning false positives can lead to unnecessary interventions and false negatives can endanger lives.
• AI platforms like Remesh and Talk to the City are being used to increase civic participation and inclusive dialogue in conflict zones, including Libya, Syria, and Yemen. While these tools can amplify marginalized voices and improve policy legitimacy, they risk excluding vulnerable groups if not supported by in-person facilitation and community partnerships.
• In humanitarian contexts, AI tools such as DISHA (led by UN Global Pulse) combine satellite imagery, drones, and participatory reports to create real-time conflict and damage maps, which in pilot studies reduced damage analysis times by a factor of six. Despite this efficiency, there is a severe risk that geospatial data could be misused to target civilians or repress dissidents.
• AI can be designed to counter the polarizing nature of social media engagement by prioritizing thoughtful comments and personal narratives over sensationalism, using tools like the Perspective API and Phoenix. These pro-social algorithms aim to increase exposure to diverse perspectives and reduce harmful disinformation to restore social cohesion.
• Effective AI deployment in fragile contexts requires strict ethical and legal oversight to prevent algorithmic discrimination, predictive policing, and the spread of 'hallucinated' or distorted facts by generative models. The document recommends investing in small-scale, context-specific models and ensuring transparency in training data and accountability mechanisms.

APA

International Panel on the Information Environment (2025). L'intelligence artificielle au service de la consolidation de la paix. https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/6909ee3c4ede2b0eeaf7b80b_Final%20FR%20AIPB%20SFP.pdf

Chicago

International Panel on the Information Environment. L'intelligence artificielle au service de la consolidation de la paix. 2025. https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/6909ee3c4ede2b0eeaf7b80b_Final%20FR%20AIPB%20SFP.pdf.

Wikipedia

{{cite report |author=International Panel on the Information Environment |title=L'intelligence artificielle au service de la consolidation de la paix |date=2025 |url=https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/6909ee3c4ede2b0eeaf7b80b_Final%20FR%20AIPB%20SFP.pdf |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{internationalpanelontheinformationenvironment2025lintelligence, author = {{International Panel on the Information Environment}}, title = {{L'intelligence artificielle au service de la consolidation de la paix}}, institution = {International Panel on the Information Environment}, year = {2025}, url = {https://cdn.prod.website-files.com/643ecb10be528d2c1da863cb/6909ee3c4ede2b0eeaf7b80b_Final%20FR%20AIPB%20SFP.pdf}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }