Abstract: What are AI-generated images pictures of? How do these statistical 
visualizations of text and image data enter into the discourse on the relationship between text and image as representative forms? This talk looks at visual generative AI on a theoretical level to understand how text and image interact and interrelate within CLIP-based AI models (contrastive language-image pre-training) such as DALL-E, Midjourney, and Stable Diffusion. Approaching this topic first from the perspective of computer vision and object recognition, I ask the question: what separates text and image or what brings them together? I address this question through an investigation and analysis of representation, using examples of AI-generated imagery and from computer vision tasks. At stake in these questions is the unraveling of popular assumptions around the truth and naturalness of photographic and other forms of visual representation.
Amanda Wasielewski is Associate Senior Lecturer of Digital Humanities and Associate Professor (Docent) of Art History in the Department of ALM (Archives, Libraries, Museums) at Uppsala University. Her research focuses on the use of artificial intelligence tools to study and create art and images. Wasielewski is the author of three monographs including Computational Formalism: Art History and Machine Learning (2023). She is also co-editor of the volume Critical Digital Art History (2024) and has recently published the article “Unnatural Images: On AI-Generated Photographs” (Critical Inquiry, Autumn 2024). Her fourth monograph, Digital Photography After AI is forthcoming in 2026.