Paper details
| Paper authors | Gülsüm Özkaya |
| In panel on | AI solutions for humanitarian challenges |
| Paper presenter(s) will be presenting |
In-Person / |
Abstract
As AI-generated visuals become increasingly common in humanitarian communication, critical questions arise regarding their ethical, emotional, and representational impact on crisis-affected communities. This paper builds on previous qualitative research I conducted at Galatasaray University, which examined AI adoption in Turkish civil society organisations. While AI tools such as ChatGPT and Midjourney are increasingly used for strategic planning and content production, the implications of AI-generated visuals—especially in representing vulnerable populations—remain underexplored.
Traditionally, humanitarian visuals have relied on photographs from crisis zones, often used without full consent. In the Syrian context, many individuals have encountered their most vulnerable moments circulating online. AI-generated visuals are seen by some NGOs as an alternative that avoids further exploitation, yet critical concerns remain: Do affected communities feel authentically represented by AI images? Does AI protect dignity or risk abstraction and dehumanisation?
To address these questions, this study draws on semi-structured interviews with Syrian social media users (aged between 18-40) affected by conflict, focusing on their perceptions of AI-generated versus traditional humanitarian imagery. The data were analysed using thematic analysis. Findings show a diversity of opinions: some participants preferred AI visuals to avoid personal exposure, while others felt such imagery diminished the seriousness of their lived experiences.
This paper argues for the inclusion of affected communities in developing ethical visual standards and calls on humanitarian actors to reassess visual practices in the age of AI.
