Widespread Gender and Pronoun Bias in Moral Judgments Across LLMs

TL;DR

This study reveals that large language models exhibit significant gender and pronoun biases in moral judgments, favoring non-binary and second-person forms, which highlights the need for fairness interventions.

Contribution

It systematically investigates how grammatical and demographic markers influence LLM moral judgments, revealing pervasive biases across multiple models.

Findings

Singular and third-person sentences are judged as more fair.

Second-person sentences are penalized in fairness judgments.

Gender markers show strong bias, favoring non-binary and disfavoring male subjects.

Abstract

Large language models (LLMs) are increasingly used to assess moral or ethical statements, yet their judgments may reflect social and linguistic biases. This work presents a controlled, sentence-level study of how grammatical person, number, and gender markers influence LLM moral classifications of fairness. Starting from 550 balanced base sentences from the ETHICS dataset, we generated 26 counterfactual variants per item, systematically varying pronouns and demographic markers to yield 14,850 semantically equivalent sentences. We evaluated six model families (Grok, GPT, LLaMA, Gemma, DeepSeek, and Mistral), and measured fairness judgments and inter-group disparities using Statistical Parity Difference (SPD). Results show statistically significant biases: sentences written in the singular form and third person are more often judged as "fair'', while those in the second person are penalized. Gender markers produce the strongest effects, with non-binary subjects consistently favored and male subjects disfavored. We conjecture that these patterns reflect distributional and alignment biases learned during training, emphasizing the need for targeted fairness interventions in moral LLM applications.