Maria-Teresa De Rosa-Palmini
Short Bio

Since February 2024, Maria-Teresa is a PhD Student at the Digital Society Initiative of the University of Zurich. She holds a Bachelor's degree in English Language and Literature from the National and Kapodistrian University of Athens and a Master's degree in Computational Linguistics from the University of Konstanz. Her current research, which is embedded within the project The Canon of Latent Spaces: How Large AI Models Encode Art and Culture under the guidance of Dr. Eva Cetinić, lies at the intersection of human-centered AI, algorithmic fairness, and cultural representation. She investigates how multimodal models, particularly text-to-image systems, encode, distort, or constrain our understanding of culture, history, and creativity. By developing systematic and scalable evaluation methods, her work uncovers representational biases in these models, tracing their roots to both training data and patterns of user interaction. More broadly, her research advances fairness in multimodal learning and supports the responsible deployment of generative AI technologies.
Research Interests
- Data-Centric AI
- AI Fairness
- Multimodal Models
- Cultural Bias
- Representation and Equity in Generative AI
Synthetic History: Evaluating Visual Representations of the Past in Diffusion Models (2025). Maria-Teresa De Rosa Palmini and Eva Cetinic. Under Review at NeurIPS 2025. Preprint available on: ArXiv
Exploring Language Patterns of Prompts in Text-to-Image Generation and Their Impact on Visual Diversity (2025). Maria-Teresa De Rosa Palmini and Eva Cetinic. ACM Conference on Fairness, Accountability, and Transparency (FAccT 2025). Available on: ArXiv , Proceedings
- Civiverse: A Dataset for Analyzing User Engagement with Open-Source Text-to-Image Models (2024). Maria-Teresa De Rosa Palmini, Laura Wagner, Eva Cetinic. Critical evaluation of generative models and their impact on society (CEGIS) workshop at European Conference of Computer Vision (ECCV 2024). Available on: ArXiv , Proceedings