FULL PDF
Abstract
This essay wonders aloud about the limitations of generative AI imagery in relationship to descriptive language. It looks at the questionably ekphrastic nature of generative AI and asks whether models and platforms such as Dall-E 2 spur creativity or constrain it by the underlying and opaque way the models use literal descriptive language.
DOI: https://doi.org/10.11588/dahj.2025.10.101535
AuthorS
Benjamin Zweig
is Project Manager for Digital Projects at Columbia University Libraries and Visiting Assistant Professor at the Pratt School of Information. He was previously the Digital Projects Coordinator at the National Gallery of Art and Research Associate for Digital Art History at the Center for Advanced Study in the Visual Arts. A medievalist by training, he received his Ph.D. in the history of art from Boston University, a M.A. in art history from Tufts University, and a B.F.A. in studio art from Massachusetts College of Art and Design. He has published widely on digital art history and medieval art, and has been the recipient of awards from he has been the recipient of awards from Boston University, the Kress Foundation, the Society for the Advancement of Scandinavian Study, and the Fulbright program.