What are chatbots' stereotypes about? A data-driven analysis of large language models' content associations with social categories

Abstract

This paper presents a data-driven analysis examining the content associations that large language models have developed with various social categories, revealing the stereotypes embedded within chatbot systems.

Publication
AAAI/ACM Conference on Artificial Intelligence, Ethics, and Society

This work investigates the stereotypes embedded in large language models by analyzing their content associations with different social categories through a comprehensive data-driven approach.

Gandalf Nicolas
Gandalf Nicolas
Assistant Professor of Psychology