Algorithmic Fairness in NLP: Persona-Infused LLMs for Human-Centric Hate Speech Detection

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Gajewska, Ewelina, Derbent, Arda, Chudziak, Jaroslaw A, Budzynska, Katarzyna
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908605930799104
author Gajewska, Ewelina
Derbent, Arda
Chudziak, Jaroslaw A
Budzynska, Katarzyna
author_facet Gajewska, Ewelina
Derbent, Arda
Chudziak, Jaroslaw A
Budzynska, Katarzyna
contents In this paper, we investigate how personalising Large Language Models (Persona-LLMs) with annotator personas affects their sensitivity to hate speech, particularly regarding biases linked to shared or differing identities between annotators and targets. To this end, we employ Google's Gemini and OpenAI's GPT-4.1-mini models and two persona-prompting methods: shallow persona prompting and a deeply contextualised persona development based on Retrieval-Augmented Generation (RAG) to incorporate richer persona profiles. We analyse the impact of using in-group and out-group annotator personas on the models' detection performance and fairness across diverse social groups. This work bridges psychological insights on group identity with advanced NLP techniques, demonstrating that incorporating socio-demographic attributes into LLMs can address bias in automated hate speech detection. Our results highlight both the potential and limitations of persona-based approaches in reducing bias, offering valuable insights for developing more equitable hate speech detection systems.
format Preprint
id arxiv_https___arxiv_org_abs_2510_19331
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Algorithmic Fairness in NLP: Persona-Infused LLMs for Human-Centric Hate Speech Detection
Gajewska, Ewelina
Derbent, Arda
Chudziak, Jaroslaw A
Budzynska, Katarzyna
Computation and Language
Computers and Society
In this paper, we investigate how personalising Large Language Models (Persona-LLMs) with annotator personas affects their sensitivity to hate speech, particularly regarding biases linked to shared or differing identities between annotators and targets. To this end, we employ Google's Gemini and OpenAI's GPT-4.1-mini models and two persona-prompting methods: shallow persona prompting and a deeply contextualised persona development based on Retrieval-Augmented Generation (RAG) to incorporate richer persona profiles. We analyse the impact of using in-group and out-group annotator personas on the models' detection performance and fairness across diverse social groups. This work bridges psychological insights on group identity with advanced NLP techniques, demonstrating that incorporating socio-demographic attributes into LLMs can address bias in automated hate speech detection. Our results highlight both the potential and limitations of persona-based approaches in reducing bias, offering valuable insights for developing more equitable hate speech detection systems.
title Algorithmic Fairness in NLP: Persona-Infused LLMs for Human-Centric Hate Speech Detection
topic Computation and Language
Computers and Society
url https://arxiv.org/abs/2510.19331