Year: 2025
Type: conference-paper
Abstract: The tendency for Natural Language Processing (NLP) technologies to reproduce stereotypical associations, such as associating Black people with criminality or women with care professions, is a site of major concern and, therefore, much study.Stereotyping is a powerful tool of oppression, but the social and linguistic mechanisms behind it are largely ignored in the NLP field.Thus, we fail to effectively challenge stereotypes and the power asymmetries they reinforce.This opinion paper problematizes several common aspects of current work addressing stereotyping in NLP, and offers practicable suggestions for potential forward directions.
Author: Hannah Devinney
Institution: Linköping University
Language: English
Subfield: Safety Research
Field: Social Sciences
Domain: Social Sciences
SDG: Reduced inequalities
Open Access status: hybrid
OpenAlex ID: W4412944386