From responsible AI to humanitarian impact: designing digital solutions for refugee response
As humanitarian crises grow in scale and complexity, organizations like UNHCR increasingly rely on data, digital technologies, and artificial intelligence to support refugee response. But what does it actually mean to use AI in humanitarian operations, and what are the opportunities and risks involved? Drawing on real UNHCR projects, this session will explore how data and AI are being applied to address operational challenges, improve decision-making, and support affected populations. Through practical examples and lessons learned, participants will gain insight into the technical, ethical, and operational considerations of deploying AI in humanitarian settings, where innovation must be balanced with accountability, transparency, and human impact.
Slides
Data at work for vulnerable lives
What does AI adoption look like in a bustling market in the Global South? How are new technologies and richer datasets expanding our potential to study and support vulnerable populations around the globe? And what are the challenges of this new era? I will touch on these and more, covering stories from my research in the humanitarian sector. I will start in Gaza and the West Bank where, by linking geocoded conflict-event data with mental health screenings, we were able to show just how closely violence tracks with depression, and put a number on the economic toll it takes on an already fragile population. From there I will move to Jordan, where we built a Climate Vulnerability Index to help identify which refugee households are most exposed to climate shocks and least equipped to cope with them. I will then close in rural Kenya, with an ongoing project where large language models are helping us listen to and decode local radio broadcasts at scale, turning an entirely new kind of data into a tool for tracking how misinformation actually spreads through communities.
Slides