PulseAugur
EN
LIVE 11:21:28

Study finds LLMs reproduce racial stereotypes in text annotation

By PulseAugur Editorial · [1 sources] ·

A withdrawn research paper found that large language models (LLMs) reproduce racial stereotypes when used for text annotation. Across 19 LLMs and over 4 million annotation judgments, the study revealed that names associated with Black individuals were rated as more aggressive, while those linked to Asian individuals were perceived as more intelligent but less sociable. The research also indicated that texts in African American Vernacular English were judged as less professional and more toxic compared to Standard American English. AI

IMPACT Highlights the risk of embedding societal biases into AI-driven annotation systems, potentially impacting research, content moderation, and hiring.

RANK_REASON Research paper on LLM bias. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Study finds LLMs reproduce racial stereotypes in text annotation

COVERAGE [1]

  1. arXiv cs.AI TIER_1 English(EN) · Petter T\"ornberg ·

    Large Language Models Reproduce Racial Stereotypes When Used for Text Annotation

    arXiv:2603.13891v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used for automated text annotation in tasks ranging from academic research to content moderation and hiring. Across 19 LLMs and two experiments totaling more than 4 million ann…

RELATED ENTITIES

RELATED TOPICS