ArabDiscrim: A Decade-Long Arabic Facebook Corpus on Racism and Discrimination

2026-05-21 · Source: Computation and Language · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Advanced, quick

Summary

ArabDiscrim is a newly released, decade-long lexical resource and corpus comprising 293,000 public Arabic Facebook posts collected between 2014 and 2024, specifically focusing on discussions of racism and discrimination. Unlike existing datasets primarily derived from Twitter, ArabDiscrim uniquely integrates platform-native engagement signals such as reactions, shares, comments, and page metadata, allowing for a joint analysis of language and audience response. The resource features 200 curated terms, split evenly between racism-related and discrimination-related concepts, each with morphological regex families covering over 13 inflections per lemma. It also delineates 20 distinct discrimination axes and provides explicit attribution patterns. Released under a restricted research-use license, ArabDiscrim supports advanced research in weak supervision, axis-aware sampling, and platform ecology, aiming to build a foundation for fairness-oriented, platform-aware Arabic Natural Language Processing.

Key takeaway

For research scientists developing fairness-oriented Arabic NLP models, ArabDiscrim offers a critical resource to move beyond Twitter-centric data. You should integrate this corpus to analyze discrimination with platform-native engagement signals, enabling more ecologically valid and nuanced understanding of online hate speech. Leverage its 20 discrimination axes and morphological depth for robust model training and evaluation, ensuring your systems are sensitive to diverse forms of identity-based unequal treatment.

Key insights

ArabDiscrim offers a unique, decade-long Arabic Facebook corpus for racism and discrimination research, integrating platform engagement signals.

Principles

• Platform-native engagement signals enhance language analysis.
• Morphological regex families improve lexical resource depth.
• Ethical compliance requires restricted data licenses.

Method

The corpus creation involved curating 200 terms with morphological regex families and identifying 20 discrimination axes from 293K public Arabic Facebook posts (2014-2024).

In practice

• Use for weak supervision in Arabic NLP.
• Enable axis-aware sampling for fairness research.
• Analyze platform ecology and audience response.

Topics

• Arabic NLP
• Discrimination Detection
• Facebook Data
• Corpus Linguistics
• Platform Ecology
• Fairness in AI

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• ArabDiscrim: A Decade-Long Arabic Facebook Corpus on Racism and Discrimination — ArabDiscrim bridges lexical depth and ecological validity for fairness-oriented, platform-aware Arabic NLP.
• AmchiBias: Measuring Stereotypical Bias in Goan Identity Groups with a Minimal Pair Dataset in English and Konkani — The AmchiBias benchmark reveals significant socio-cultural bias gaps in multilingual NLP for low-resource, hyperlocal communities like Goa.
• CATCH-ME if you RAG: a dataset of Contextually Annotated multi-Turn Counterspeech against Hate and Misinformation Exchanges — CATCH-ME provides a multilingual, multi-turn, knowledge-grounded dataset for countering intertwined hate speech and misinformation.
• CRED-1: Building an Open Domain Credibility Dataset for Misinformation Pre-Bunking — Multi-signal credibility scoring offers a nuanced, dynamic, and transparent alternative to binary domain labeling for misinformation pre-bunking.
• IndoBias: A Dual Track Culturally Grounded Benchmark for LLMs Bias Evaluation in Indonesian Languages — Culturally-grounded benchmarks are crucial for evaluating LLM bias in diverse, multilingual contexts like Indonesia.
• Audience Engagement with Arabic Women's Social Empowerment and Wellbeing: A Decadal Corpus — The corpus provides a decadal view of Arabic women's social empowerment discourse on Facebook.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computation and Language.