A Close Reading Approach to Gender Narrative Biases in AI-Generated Stories

TL;DR

This study investigates gender biases in AI-generated stories using a close reading method, revealing persistent implicit biases across different AI models and emphasizing the need for multi-level bias assessment.

Contribution

It introduces a novel interpretative approach to analyze gender biases in AI storytelling, combining literary analysis with bias detection.

Findings

Biases persist across AI models including ChatGPT, Gemini, and Claude.

Implicit gender biases are prevalent in character descriptions and plot development.

Multi-level bias assessment is crucial for understanding AI-generated narrative biases.

Abstract

The paper explores the study of gender-based narrative biases in stories generated by ChatGPT, Gemini, and Claude. The prompt design draws on Propp's character classifications and Freytag's narrative structure. The stories are analyzed through a close reading approach, with particular attention to adherence to the prompt, gender distribution of characters, physical and psychological descriptions, actions, and finally, plot development and character relationships. The results reveal the persistence of biases - especially implicit ones - in the generated stories and highlight the importance of assessing biases at multiple levels using an interpretative approach.