▶
Out of the Box: High-Risk by Default? Humanitarian AI Under Pressure
Artificial intelligence (AI) is no longer a distant prospect for humanitarian action. It is already in use. From anticipatory action to biometrics, predictive analytics and data analysis, both traditional and generative AI are increasingly becoming part of how humanitarian actors operate. However, organisational and sector-wide policies, capacities and oversight remain lacking. This unpreparedness creates capability gaps, ethical dilemmas, liability risks and uncertainties as humanitarian actors navigate high-stakes decisions in fragile contexts. At the same time, there is no coordinated space for practitioners to share knowledge and develop guidance across operational, technical and legal functions.
This Out of the Box session presents new studies on humanitarian innovation and the use and procurement of AI tools, highlighting potential harms for affected populations and the responsibilities of humanitarian actors, convening a cross-functional dialogue on how to move from AI awareness to collective action and AI readiness.
It presents key findings of the new CHA research paper “Balancing innovation, efficiency, and principled humanitarian action”, authored by Andrea Düchting and Darina Pellowska and published in August 2025 (https://www.chaberlin.org/en/publications/balancing-innovationefficiency-and-principled-humanitarian-action-2/). The event features inputs by Giulio Coppi, Senior Humanitarian Officer at Access Now; Pierrick Devidal, Humanitarian Policy Adviser at the ICRC and Stefan Delfs, Head of Division S07 at the GFFO.
____________________________________________________________
To keep up to date with the work of the Centre for Humanitarian Action (CHA), visit our website https://www.chaberlin.org/
And follow us on:
Bluesky https://bsky.app/profile/chaberlin.bsky.social
Facebook https://www.facebook.com/chaberlin.org
LinkedIn https://www.linkedin.com/company/centre-for-humanitarian-action/
