ExTax: Explainable Disinformation Detection via Persuasion, Emotion, and Narrative Role Taxonomies

2026-05-26

AI Summary

This study addresses the limitations of existing misinformation detection approaches, which struggle to model multidimensional manipulative intent and lack interpretability. The authors propose the first interpretable 17-dimensional classification framework that integrates persuasive strategies, emotional manipulation, and narrative roles. Leveraging large language models, the method extracts fine-grained attributes and enhances robustness and auditability through entropy-driven dynamic label smoothing and a heterogeneous multi-head attention mechanism. Evaluated across five cross-domain, cross-genre benchmarks, the approach achieves a Macro F1 score of 0.8456, significantly outperforming current deep learning and large-model baselines, while maintaining stable performance even under severe class imbalance.

Abstract

The democratization of LLMs has accelerated the generation and circulation of highly fluent disinformation, making traditional syntax-semantic verification increasingly insufficient. Such deception rarely relies solely on surface-level falsity; instead, it often combines persuasive rhetoric, emotional manipulation, and narrative role construction to influence readers' interpretations through multiple cognitive pathways. However, existing detectors typically emphasize isolated signals -- such as syntax, external knowledge, persuasion, or affective cues -- and therefore struggle to capture the multi-faceted manipulative intents underlying disinformation or provide human-auditable explanations. To address this gap, we present \textbf{ExTax}, a taxonomy-aligned framework for explainable disinformation detection. ExTax unifies persuasive rhetoric, emotional manipulation, and narrative roles into a 17-dimensional taxonomic space, covering 6 persuasive-rhetoric strategies, 5 emotional-manipulation methods, and 6 narrative-role categories. It elicits attributes from multiple frontier LLMs, reconciles their disagreements through Entropy-driven Dynamic Label Smoothing, and fuses the resulting taxonomic representations with contextual encodings via Heterogeneous Multi-Head Attention, grounding each prediction in an interpretable manipulation profile. Across five cross-domain and cross-genre benchmarks, ExTax achieves an overall Macro $F_1$ of $0.8456$, outperforming state-of-the-art deep learning and LLM-based baselines. It also remains robust under severe genre imbalance, where the strongest deep baseline degrades from $0.9454$ to $0.6194$.

Problem

• disinformation detection
• explainable AI
• persuasive rhetoric
• emotional manipulation
• narrative roles

Innovation

• Explainable Disinformation Detection
• Taxonomy-Aligned Framework
• Persuasion-Emotion-Narrative Integration
• Entropy-driven Dynamic Label Smoothing
• Heterogeneous Multi-Head Attention