DariMis: Harm-Aware Modeling for Dari Misinformation Detection on YouTube

TL;DR

This paper introduces DariMis, a new annotated dataset for Dari-language YouTube videos, and proposes a pair-input encoding method that improves misinformation detection accuracy by modeling semantic relationships between video titles and descriptions.

Contribution

It presents the first Dari misinformation dataset, analyzes the coupling of misinformation and harm levels, and introduces a pair-input encoding strategy that enhances detection performance.

Findings

Over half of misinformation videos carry medium or high harm potential.

Pair-input encoding improves misinformation recall by 7 percentage points.

ParsBERT outperforms XLM-RoBERTa-base on Dari misinformation detection.

Abstract

Dari, the primary language of Afghanistan, is spoken by tens of millions of people yet remains largely absent from the misinformation detection literature. We address this gap with DariMis, the first manually annotated dataset of 9,224 Dari-language YouTube videos, labeled across two dimensions: Information Type (Misinformation, Partly True, True) and Harm Level (Low, Medium, High). A central empirical finding is that these dimensions are structurally coupled, not independent: 55.9 percent of Misinformation carries at least Medium harm potential, compared with only 1.0 percent of True content. This enables Information Type classifiers to function as implicit harm-triage filters in content moderation pipelines. We further propose a pair-input encoding strategy that represents the video title and description as separate BERT segment inputs, explicitly modeling the semantic relationship between headline claims and body content, a key signal of misleading information. An ablation study against single-field concatenation shows that pair-input encoding yields a 7.0 percentage point gain in Misinformation recall (60.1 percent to 67.1 percent), the safety-critical minority class, despite modest overall macro F1 differences (0.09 percentage points). We benchmark a Dari/Farsi-specialized model (ParsBERT) against XLM-RoBERTa-base; ParsBERT achieves the best test performance with accuracy of 76.60 percent and macro F1 of 72.77 percent. Bootstrap 95 percent confidence intervals are reported for all metrics, and we discuss both the practical significance and statistical limitations of the results.