From flood prediction systems to optimizing resource distribution, humanitarian programs increasingly depend on real-time and historical data analysis. 510 is an initiative by the Netherlands Red Cross that harnesses digital and data-driven solutions to support various humanitarian projects worldwide. In this context, Raoul Brigola, an MSc student from Utrecht University, collaborated with the 510 team to address a significant challenge in humanitarian programs: understanding people’s personal values in crisis situations through data analysis. Under the guidance of HI members Davide Dall’Anna and Pinar Yolum, Raoul focused on processing chat messages from Telegram groups used by Ukrainian refugees and internally displaced people to gain deeper insights into their values and needs. 510 had already utilized automated tools to extract discussion topics from these Telegram groups. However, identifying latent personal values, such as fairness and self-determination, often requires more nuanced human interpretation beyond what automated tools can achieve.

To tackle this, Raoul proposed VIVE (Value Identification and Value Extraction), a Hybrid Intelligence method designed to identify and extract personal values from text. VIVE builds on Axies, another Hybrid Intelligence tool developed by HI member Enrico Liscio, which uses NLP techniques to guide humans in identifying relevant values in large datasets. By leveraging Large Language Models, VIVE then quantifies the prevalence of the identified values in the dataset. In collaboration with 510 analysts, Raoul applied VIVE to analyze the values expressed by the Ukrainian refugees. While expected values such as health and financial concerns were prominent, VIVE’s Hybrid Intelligence approach also surfaced less obvious values that had not been identified through the previously used automated methods, like the feeling of being unwelcomed in host countries. This highlights how combining human judgment with AI scalability is instrumental in exploring refugees’ values and enhancing humanitarian efforts.