Wikigender: A Machine Learning Model to Detect Gender Bias in Wikipedia

TL;DR

This paper presents a machine learning approach to detect gender bias in Wikipedia biographies by analyzing adjectives and topics, revealing gender stereotypes and subjective portrayals.

Contribution

It introduces a novel ML model that identifies gender bias in Wikipedia text, focusing on adjectives and topics to uncover stereotypes and subjective language.

Findings

Women are described with more subjective adjectives.

Women are associated with family topics.

Men are linked to business and sports.

Abstract

The way Wikipedia's contributors think can influence how they describe individuals resulting in a bias based on gender. We use a machine learning model to prove that there is a difference in how women and men are portrayed on Wikipedia. Additionally, we use the results of the model to obtain which words create bias in the overview of the biographies of the English Wikipedia. Using only adjectives as input to the model, we show that the adjectives used to portray women have a higher subjectivity than the ones used to describe men. Extracting topics from the overview using nouns and adjectives as input to the model, we obtain that women are related to family while men are related to business and sports.