Title: BOUTEF: A Multilingual Corpus for FakeNews in North Africa -- Language as a Weapon

Abstract

The proliferation of misinformation on social platforms poses a significant challenge, especially within multilingual and under-resourced regions like North Africa. This study presents BOUTEF, a large-scale multilingual dataset created to investigate the dissemination, features, and effects of fake news in Algeria and Tunisia. BOUTEF comprises three distinct elements: fabricated stories, authentic narratives, and the accompanying user comments, supplemented by verified debunking data. The resource encompasses a diverse array of linguistic forms, including Modern Standard Arabic (MSA), Algerian and Tunisian dialects, Arabizi, French, English, and various code-switched varieties.

Leveraging this dataset, we perform a thorough empirical analysis that merges quantitative and qualitative methodologies. Our investigation delves into thematic distributions, linguistic and rhetorical tactics, sentiment trends, and the dynamics of social engagement. Statistical evaluations uncover substantial links between specific thematic categories and the truthfulness of messages, as well as a pronounced correlation between user interaction levels and the prominence of false content.

Our results indicate that fake news predominantly utilizes emotionally resonant narratives, sensationalist framing, and mixed linguistic practices to boost virality and audience involvement. Conversely, debunking materials tend to employ a more objective and verification-focused tone. Additionally, a cross-country comparison between Algeria and Tunisia reveals both common patterns and unique traits influenced by their respective sociopolitical landscapes. These findings underscore the importance of informal language usage in both the spread and reception of misinformation. By offering a comprehensive, annotated, and open-access dataset, this research aims to advance studies on fake news detection, low-resource language processing, and the comprehension of information disorders in complex linguistic settings.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC