Summary of Human and Llm Biases in Hate Speech Annotations: a Socio-demographic Analysis Of Annotators and Targets, by Tommaso Giorgi et al.
Human and LLM Biases in Hate Speech Annotations: A Socio-Demographic Analysis of Annotators and Targets
by Tommaso Giorgi, Lorenzo Cima, Tiziano Fagni, Marco Avvenuti, Stefano Cresci
First submitted to arxiv on: 10 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper explores the relationship between the characteristics of individuals who label hate speech data and the attributes of the target of that hate speech. It uses an extensive dataset with socio-demographic information to uncover biases in human labeling, finding widespread biases that vary in intensity and prevalence. The findings are compared to those of persona-based language models (LLMs), which exhibit different biases than humans. The research aims to provide new insights into designing AI-driven hate speech detection systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how people’s characteristics affect their judgment about what is considered hateful online. It uses a big dataset with lots of information about the people who labeled the data and the things they were labeling as hate speech or not. The research found that most people have biases when it comes to deciding what is hateful, and these biases are different depending on things like age, gender, and race. The study also compares how AI machines make decisions with human judgments, showing that while both can be biased, the biases are different. |
