DH Lunch: ImageTrain - Uncovering generative AI's hidden assumptions
In this DH-Lunch, you will learn how AI-generated images are shaping today’s visual culture, while the assumptions embedded in these models through their training data often remain hidden.
ImageTrain by the Department of Art and Media Studies (IKM) at NTNU, in cooperation with D-lab at NTNU (external link) and Prof. Moritz Schwind (Professor of Computational Design for Future Systems, TH Nürnberg) makes them visible by provoking and documenting productive failure in practice-based experiments that are later intended for our students as well, to repeat, test, and experiment with own small mobile phone datasets. Handcrafted drawings within an animated video essay by a professional illustrator / artist will later also “re-draw” our AI image generation in a video essay for easier understanding, making the cultural dimensions of AI legible to researchers, artists, cultural practitioners, and students.
About the lecturer
Aurora Hoel (external link) is Professor of Media Studies and Visual Culture in the Department of Art and Media studies at NTNU. Together with Lukas Wilde, she leads the research group AI Media (external link) which investigates the transformative effects of machine learning and generative AI on media and society.