Nouns So Abstract - Part One, Birth: Marginalized Images in Generative AI

Abstract

“We explained her diagnosis to her, about the tumors, the cells, metastasis. Nouns so abstract that we might as well have been describing witchcraft.” -Ocean Vuong, On Earth We’re Briefly Gorgeous Inspired by Vuong’s metaphor of medical terminologies’ equivalence to witchcraft and the inaccessibility of language through intergenerational assimilation, Nouns So Abstract explores the visualization of a “Chinese Witch.” Part one, Birth, explores the designer’s experience giving birth to her first and only child, from midwife pre-natal visits, failed doula coaching, acute hospital trauma, and postpartum depression. For Birth, she has created interactive animations from collected AI still renderings of keywords (aka nouns) that include: “chinese” + “witch” + “natural” + “bonding” ++. The AI platforms are all open-access and includes DALL-E 2, Midjourney and DeepAI. The AI renderings showcase what the Internet believes is a Chinese Witch, highlighting misconstructions, demonization, and stereotypes inherent in these AI learning algorithms; and questions who owns the images of marginalized people?

Presenters

Zhifang Li
Assistant Professor, Digital Arts, La Salle University, United States

Details

Presentation Type

Poster Session

Theme

2024 Special Focus—Cultures of Transformative Design

KEYWORDS

AI, Marginalize, Chinese, Witch, Language, Video, Installation, Design, Art, Technology