Addressing power dynamics in participatory AI for crisis- affected communities (beta version)

Ariel view of Kakuma refugee camp, Kenya. Credit: FilmAid Kenya.

Without meaningful participation from crisis-affected communities, humanitarian AI risks becoming experimental and extractive, worsening inequalities rather than addressing them.

This case study from the SAFE AI Project examines real-world challenges in implementing participatory AI through a groundbreaking pilot in Kakuma Refugee Camp and Kalobeyei Settlement.

In December 2024, CDAC Network and FilmAid conducted the first documented effort to engage refugee communities in AI design and governance. The pilot revealed critical gaps between participatory AI theory and humanitarian practice, exposing power imbalances that undermine genuine community involvement.

This policy brief identifies key obstacles preventing meaningful participation and provides actionable recommendations for ethical, inclusive participatory AI in humanitarian settings. Essential reading for practitioners committed to shifting power dynamics and ensuring AI solutions serve—rather than exploit—crisis-affected communities.

This material has been funded by UK International Development from the UK government.


Next
Next

Co-design vs. User-Centred Design for AI Solutions (Beta version)