Double Delight in Geneva: AI for Social Good + Crisis Computing Workshop

From July 8–11, 2025, I (Theophilus Aidoo) had the privilege of attending two significant events in one trip: the AI for Social Good Global Summit and the Crisis Computing Workshop, both held in the beautiful city of Geneva. It was an inspiring and intellectually rich experience that brought together practitioners and researchers using AI to address societal challenges. 

At the AI for Social Good Summit, I was particularly impressed by the variety of applications showcased from agriculture and disaster response to robotics. One session that stood out for me was the Human-Centered AI for Disaster Management session. There, I learned how researchers are leveraging large language models (LLMs) to assist in the analysis and interpretation of satellite imagery. This approach not only improves the efficiency of remote assessments but also enables more human-readable and accessible communication of technical results. 

A quote that stayed with me from that session was: 

“Data must be owned by the local people.” 

This simple yet powerful statement underscores the importance of local context in data collection and analysis, especially in crisis settings. It is a reminder that any meaningful solution must reflect the realities on the ground and that local communities must have both input into and ownership of the data used to make decisions that affect them. 

Another highlight was the concluding question posed by Leonardo Milano during his keynote: 

“How might we conduct rapid estimates in the event of a disaster?” 

Although my current work focuses more on displacement estimation in crisis settings, this question resonated with me. It aligns with broader goals in my research, and I hope that my contributions particularly in satellite image analysis can support efforts to address such urgent needs in the future.

In addition to the summit, I also participated in the  Crisis Computing Workshop, which offered a more focused and collaborative environment. It was a unique opportunity to engage directly with researchers and industry professionals working at the intersection of AI, and humanitarian response in crises settings, where one could discuss their ideas and get feedback on them. My take away from this event apart from the conversations was the diversity of data sources being explored from social media activity and radio signals to satellite imagery and other crisis data sources such as ACLED. These conversations significantly expanded my thinking about what is possible when technical innovation is grounded in real-world operational needs.

Throughout both events, I appreciated the emphasis on interdisciplinary collaboration and the value of bringing together different perspectives from academia, the humanitarian sector, and technology companies. The trip left me with a deeper appreciation for context-driven AI, a renewed motivation to contribute to the space of humanitarian technology, and a sense of belonging in a global community committed to using AI for the common good.