As humanitarian actors turn to AI to improve crisis response, new risks arise, especially in conflict zones where data processing can inadvertently cause harm, increase vulnerability, and even skew the distribution of aid. Led by Dr. Tino Kreutzer, a renowned expert in AI data solutions for humanitarian settings, this session explores the critical question: How can we develop and deploy AI responsibly in humanitarian crises without increasing risk to affected communities?

With ICT4D and monitoring and evaluation experts from leading humanitarian and development organizations, participants will work through real-world decision points drawn from humanitarian emergencies. Attendees will gain insights into critical ethical and practical challenges such as how to use AI models effectively for aid allocation, how to address unclear consent, and how to ensure AI-generated recommendations are tailored to local contexts. We will explore methods for identifying risks, leveraging AI for better outcomes and greater impact, and practical next steps for designing and implementing AI tools in humanitarian crises.