# Who Speaks Matters: Analysing the Influence of the Speaker's Ethnicity on Hate Classification

arXiv: https://arxiv.org/abs/2410.20490 · HTML: https://mlsanity.com/paper/2410.20490/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 38 references that are in the map.
- First posted on arXiv 27 Oct 2024.

## Builds on (most cited first, 15 of 15)

- [Beyond Accuracy: Behavioral Testing of NLP models with CheckList](https://mlsanity.com/paper/2005.04118/) (2020)
- [HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection](https://mlsanity.com/paper/2012.10289/) (2020)
- [Should ChatGPT be Biased? Challenges and Risks of Bias in Large Language Models](https://mlsanity.com/paper/2304.03738/) (2023)
- [Towards Making the Most of ChatGPT for Machine Translation](https://mlsanity.com/paper/2303.13780/) (2023)
- [Is LLM-as-a-Judge Robust? Investigating Universal Adversarial Attacks on Zero-shot LLM Assessment](https://mlsanity.com/paper/2402.14016/) (2024)
- [A Survey of Race, Racism, and Anti-Racism in NLP](https://mlsanity.com/paper/2106.11410/) (2021)
- [Investigating Bias in LLM-Based Bias Detection: Disparities between LLMs and Human Perception](https://mlsanity.com/paper/2403.14896/) (2024)
- [Exploring Cross-Cultural Differences in English Hate Speech Annotations: From Dataset Construction to Analysis](https://mlsanity.com/paper/2308.16705/) (2023)
- [Unsupervised Discovery of Implicit Gender Bias](https://mlsanity.com/paper/2004.08361/) (2020)
- [HinGE: A Dataset for Generation and Evaluation of Code-Mixed Hinglish Text](https://mlsanity.com/paper/2107.03760/) (2021)
- [Don't Go To Extremes: Revealing the Excessive Sensitivity and Calibration Limitations of LLMs in Implicit Hate Speech Detection](https://mlsanity.com/paper/2402.11406/) (2024)
- [Hate Personified: Investigating the role of LLMs in content moderation](https://mlsanity.com/paper/2410.02657/) (2024)
- [How Hate Speech Varies by Target Identity: A Computational Analysis](https://mlsanity.com/paper/2210.10839/) (2022)
- [From Human Judgements to Predictive Models: Unravelling Acceptability in Code-Mixed Sentences](https://mlsanity.com/paper/2405.05572/) (2024)
- [Examining risks of racial biases in NLP tools for child protective services](https://mlsanity.com/paper/2305.19409/) (2023)

## Built on by (most cited first, 1 of 1)

- [Are LLMs Empathetic to All? Investigating the Influence of Multi-Demographic Personas on a Model's Empathy](https://mlsanity.com/paper/2510.10328/) (2025)

## Abstract (arXiv)

Large Language Models (LLMs) offer a lucrative promise for scalable content moderation, including hate speech detection. However, they are also known to be brittle and biased against marginalised communities and dialects. This requires their applications to high-stakes tasks like hate speech detection to be critically scrutinized. In this work, we investigate the robustness of hate speech classification using LLMs particularly when explicit and implicit markers of the speaker's ethnicity are injected into the input. For explicit markers, we inject a phrase that mentions the speaker's linguistic identity. For the implicit markers, we inject dialectal features. By analysing how frequently model outputs flip in the presence of these markers, we reveal varying degrees of brittleness across 3 LLMs and 1 LM and 5 linguistic identities. We find that the presence of implicit dialect markers in inputs causes model outputs to flip more than the presence of explicit markers. Further, the percentage of flips varies across ethnicities. Finally, we find that larger models are more robust. Our findings indicate the need for exercising caution in deploying LLMs for high-stakes tasks like hate speech detection.