Synthetic Genders: How Generative AI Reimagines Gender in Visual Culture

The Visual Imaginaries of Gender project investigates the potentials and limitations of Generative AI (GenAI) for image generation. It goes beyond the known findings of biases in who is represented in generated images by deepening our understanding of GenAI’s artistic reimaginations of gender. During this evening we explore how the different scientific and artistic research perspectives brought together in the project intertwine and collide in three talks: 

Piera Riccio will present her work through four key perspectives. First, she examines GenAI as a widespread cultural phenomenon, exploring how users attribute sexualized and emotional connotations to fictional female characters. Second, how gendered artifacts from historical artistic styles are reproduced and amplified. Third, on continuities between existing digital beautification practices and AI portrait aesthetics, where users employ generative tools to “enhance” their appearance. Finally, she reflects on the circulation of abusive depictions of women on social media, highlighting tensions between freedom of expression, content moderation, and the prevention of gender-based violence in the age of generative AI. 

Eftychia Stamkou will explore how visual culture reflects but also shapes dominant cultural narratives. Using AI to analyze a large collection of paintings and film scripts, she asks: How has the representation of female artists shifted over time and across cultures? Does a female director guarantee a female voice on screen? And are women in paintings depicted as active agents or passive muses? 

Melvin Wevers will discuss how AI models encode temporal patterns. Drawing on his research analyzing visual markers in historical photographs, he explores how AI learns associations between time periods and gender. From a historian’s perspective, he asks: what do these models reveal about how we collectively remember and visualize the past? These models don’t just generate images of the past, but actively construct which versions of gendered history become visible and reproducible.  

The evening will conclude with a panel discussion led by Maarten Wijntjes. 

Speakers 

Piera Riccio is a Postdoctoral Researcher Multimedia Analytics lab of the University of Amsterdam, working on the Visual Imaginaries of Gender project. Her research relies on techniques and practices from Computer Vision, Gender Studies and Human-Computer Interaction, with a specific focus on the intersection between AI technologies and artistic practices. In 2025, she obtained a PhD from ELLIS Alicante/University of Alicante, with a thesis titled “Human Aesthetics under the representational power of Artificial Intelligence”. 

Eftychia Stamkou is an Assistant Professor of Social and Cultural Psychology at the University of Amsterdam. Her research examines how norm violators gain power in art, business, and politics. She’s also studying the developmental roots of early cultural engagement. Her work bridges academic inquiry with real-world impact, serving as a consultant for businesses and regularly collaborating with cultural institutions, including Google Arts & Culture, Carnegie Hall, Singer Laren, and NEMO. She directs the Amsterdam Arts and Social Sciences lab at UvA and organizes the Where Art Meets Science workshop series at Institute forAdvanced Study.  

Melvin Wevers is a computational historian at the University of Amsterdam, applying multimodal machine learning to historical archives. His research spans computer vision analysis of historical visual archives, word embeddings for tracking conceptual change in newspapers, and phylogenetic diversity metrics for diagnosing archival silences. He develops methods that bridge critical archival theory with quantitative analysis to reveal how institutional systems, cataloging practices, and technological constraints systematically shape historical representation and digital accessibility. 

Nanne van Noord is Assistant Professor at the Multimedia Analytics lab of the University of Amsterdam. His research lies at the intersection of Multimodal AI and Visual Culture, with the aim of integrating equitable visual cultural understanding into AI models to bridge the gap between humanistic and algorithmic inquiry. 

Maarten Wijntjes (moderator) currently serves as Associate Professor of Visual Perception and Communication at Delft University of Technology’s Faculty of Industrial Design Engineering. As the chair of thePictorial Research Lab he investigates and teaches about three topics: Pictorial Analysis (with humans and machines), Pictorial Experience (e.g. with optical devices) and Pictorial Communication (quantifying perceptual differences between representation and reality). His expertise extends beyond the pictorial, exploring how these principles apply to the broader multi-sensory domain.