Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Biases in Generative Artificial Intelligence: A visual semiotic analysis of AI Images

Christiansen, Sofi LU (2024) KOVM12 20241
Division of Art History and Visual Studies
Abstract
This thesis focuses on images produced by Generative Artificial Intelligence and poses questions around biases and implication of these. With the emergence of AI image creation programs, the complexity of representation is brought to the surface. The main material in this thesis consists of AI-generated images and the material will be analyzed with a semiotic framework through the lens of cultural, gaze, and gender theory. This thesis confirms previous research on AI-generated images, which found that images contain biases and reinforce whiteness as a norm. It further contests previous papers on the topic by arguing that we should expect more from AI programs. Lastly, it discusses the importance of images that hinder the amplification of... (More)
This thesis focuses on images produced by Generative Artificial Intelligence and poses questions around biases and implication of these. With the emergence of AI image creation programs, the complexity of representation is brought to the surface. The main material in this thesis consists of AI-generated images and the material will be analyzed with a semiotic framework through the lens of cultural, gaze, and gender theory. This thesis confirms previous research on AI-generated images, which found that images contain biases and reinforce whiteness as a norm. It further contests previous papers on the topic by arguing that we should expect more from AI programs. Lastly, it discusses the importance of images that hinder the amplification of stereotypes and introduces the concept of Conscious Prompting, an approach for generating images for marketing purposes to bridge the gap between the humanities and the industry. (Less)
Please use this url to cite or link to this publication:
author
Christiansen, Sofi LU
supervisor
organization
course
KOVM12 20241
year
type
H2 - Master's Degree (Two Years)
subject
keywords
AI, Biases, Generative Artificial Intelligence, Representation, Stereotypes
language
English
id
9161589
date added to LUP
2024-09-24 11:49:23
date last changed
2024-09-24 11:49:23
@misc{9161589,
  abstract     = {{This thesis focuses on images produced by Generative Artificial Intelligence and poses questions around biases and implication of these. With the emergence of AI image creation programs, the complexity of representation is brought to the surface. The main material in this thesis consists of AI-generated images and the material will be analyzed with a semiotic framework through the lens of cultural, gaze, and gender theory. This thesis confirms previous research on AI-generated images, which found that images contain biases and reinforce whiteness as a norm. It further contests previous papers on the topic by arguing that we should expect more from AI programs. Lastly, it discusses the importance of images that hinder the amplification of stereotypes and introduces the concept of Conscious Prompting, an approach for generating images for marketing purposes to bridge the gap between the humanities and the industry.}},
  author       = {{Christiansen, Sofi}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Biases in Generative Artificial Intelligence: A visual semiotic analysis of AI Images}},
  year         = {{2024}},
}