CDAC Network Community of Practice

AI governance

AI is already used across humanitarian response. It affects who receives aid, what information people see and how decisions about them are made. Most of these systems have no independent check on how they work or whom they affect.

CDAC Network convenes a Community of Practice on AI governance and crisis-affected communities. It is a forum for people who use, buy, build, fund or question AI in crisis settings. It is open to anyone working on these issues, including people outside CDAC membership.

Participants use the Community of Practice to: share practical experience of AI in humanitarian programmes, including what did not work; work through governance questions with peers, such as risk assessment, supplier due diligence and how to inform affected communities; test and give feedback on SAFE AI tools and guidance; build common positions for policy and donor discussions.

The Community of Practice connects directly to the SAFE AI Initiative, which CDAC Network stewards. SAFE AI (Standards and Assurance for Ethical AI) starts with humanitarian action. It gives organisations a common way to govern and document the AI they use, and to prepare for independent verification. It was developed with The Alan Turing Institute and Humanitarian AI Advisory, with founding investment from UK International Development from the UK government. The SAFE AI Framework and tools are published and free to use.

Both put community in the loop at every stage of the AI lifecycle. People affected by crisis have a say in how AI is designed, used and governed.