Beyond the Explicit: A Bilingual Dataset for Dehumanization Detection in Social Media

TL;DR

This paper introduces a bilingual dataset for detecting subtle forms of dehumanization in social media, addressing a gap in existing research by capturing nuanced biases and providing a benchmark for future models.

Contribution

It presents a new, theory-informed bilingual dataset with extensive annotations, and demonstrates improved machine learning detection performance over existing models.

Findings

Dataset covers multiple dehumanization dimensions

Models fine-tuned on this dataset outperform state-of-the-art

Effective in zero and few-shot learning scenarios

Abstract

Digital dehumanization, although a critical issue, remains largely overlooked within the field of computational linguistics and Natural Language Processing. The prevailing approach in current research concentrating primarily on a single aspect of dehumanization that identifies overtly negative statements as its core marker. This focus, while crucial for understanding harmful online communications, inadequately addresses the broader spectrum of dehumanization. Specifically, it overlooks the subtler forms of dehumanization that, despite not being overtly offensive, still perpetuate harmful biases against marginalized groups in online interactions. These subtler forms can insidiously reinforce negative stereotypes and biases without explicit offensiveness, making them harder to detect yet equally damaging. Recognizing this gap, we use different sampling methods to collect a theory-informed bilingual dataset from Twitter and Reddit. Using crowdworkers and experts to annotate 16,000 instances on a document- and span-level, we show that our dataset covers the different dimensions of dehumanization. This dataset serves as both a training resource for machine learning models and a benchmark for evaluating future dehumanization detection techniques. To demonstrate its effectiveness, we fine-tune ML models on this dataset, achieving performance that surpasses state-of-the-art models in zero and few-shot in-context settings.