Work

Characterizing non-binary French: A first step towards debiasing gender inference

Year: 2025

Type: conference-paper

Abstract: This paper addresses a bias of gender inference systems: their binary nature.Based on the observation that, for French, systems based on pattern-matching of grammatical gender markers in "I am" expressions perform better than machine-learning approaches (Ciot et al., 2013), we examine the use of grammatical gender by non-binary individuals.We describe the construction of a corpus of texts produced by non-binary authors on Reddit, (formely) Twitter and three forums.Our linguistic analysis shows three main patterns of use: authors who use non-binary markers, authors who consistently use one grammatical gender, and authors who use both feminine and masculine markers.Using this knowledge, we make proposals for the improvements of existing gender inference systems based on grammatical gender.

Source: Workshop on Gender Bias in Natural Language Processing (GeBNLP)

DOI: 10.18653/v1/2025.gebnlp-1.16

Authors: Marie Flesch, Heather Burnett

Institutions: Université Paris Cité, Laboratoire de Linguistique Formelle, Analyse et Traitement Informatique de la Langue Française, Université de Lorraine

Language: English

Topic: Authorship Attribution and Profiling

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDGs: Reduced inequalities, Gender equality

Open Access status: hybrid

Funders: European Commission, Agence Nationale de la Recherche