Date of Award

5-2026

Degree Type

Creative Project

Degree Name

Master of Science (MS)

Department

History

Committee Chair(s)

Afsane Rezaei (Committee Chair)

Committee

Afsane Rezaei

Committee

Lynne S. McNeill

Committee

Sharad Jones

Abstract

When asked to generate a monster, AI image generators consistently produce masculine or genderless creatures and never female ones. This thesis documents that finding through a year-long experiment across seven generators that produced 577 monster images, not one of which depicted a female monster. Female monstrosity has deep roots in folklore and cultural tradition, yet it is entirely absent from these systems' neutral outputs. When specifically prompted, the generators handled female monstrosity in wildly different ways, ranging from hypersexualized figures to a dragon wearing a dress, while male monsters looked nearly identical across all platforms. The inconsistency between generators indicates that the bias is not just a product of the source data but is influenced by how each system is trained, which means it can be identified and corrected. As AI-generated content feeds back into future training data, these patterns risk reinforcing themselves over time. This thesis draws on folklore studies and gender scholarship to argue that the cultural biases encoded within AI systems are a matter of real societal concern and warrant attention from both researchers and developers.

Comments

This work is comprised of two files. The first is the main thesis and the second is an Appendix which contains the full dataset of generated monster imagery collected and referenced in the thesis itself.

Download

Additional Files

Nikki Christensen Thesis Appendix A and B.docx (378364 kB)
Appendix A and B with Monster Images

Share

COinS

Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .