Detecting misinformation through Framing Theory: the Frame Element-based Model

TL;DR

The paper addresses misinformation that arises from framing accurate facts, proposing a Frame Element-based Model (FEM) that integrates framing theory into neural architectures. By extracting four frame elements—Problem Definition, Causal Interpretation, Moral Evaluation, and Treatment Recommendation—via an LLM-based frame extractor and encoding them alongside the article, FEM uses a Bi-LSTM and a softmax classifier to detect framed misinformation, with a loss that combines cross-entropy and regularization. Empirical results across four datasets show that FEM with both text and frame elements outperforms strong baselines, and ablation studies reveal the critical role of certain frame elements (notably Problem Definition and Moral Evaluation) while highlighting topic-dependent effects. The work demonstrates that incorporating framing structure into misinformation detection enhances accuracy and reveals nuanced, topic-sensitive patterns, suggesting practical implications for more trustworthy AI systems and targeted misinformation mitigation strategies. The approach combines formal framing definitions, dataset augmentation through frame manipulation, and rigorous evaluation to advance understanding of how narrative framing drives misinformation.

Abstract

In this paper, we delve into the rapidly evolving challenge of misinformation detection, with a specific focus on the nuanced manipulation of narrative frames - an under-explored area within the AI community. The potential for Generative AI models to generate misleading narratives underscores the urgency of this problem. Drawing from communication and framing theories, we posit that the presentation or 'framing' of accurate information can dramatically alter its interpretation, potentially leading to misinformation. We highlight this issue through real-world examples, demonstrating how shifts in narrative frames can transmute fact-based information into misinformation. To tackle this challenge, we propose an innovative approach leveraging the power of pre-trained Large Language Models and deep neural networks to detect misinformation originating from accurate facts portrayed under different frames. These advanced AI techniques offer unprecedented capabilities in identifying complex patterns within unstructured data critical for examining the subtleties of narrative frames. The objective of this paper is to bridge a significant research gap in the AI domain, providing valuable insights and methodologies for tackling framing-induced misinformation, thus contributing to the advancement of responsible and trustworthy AI technologies. Several experiments are intensively conducted and experimental results explicitly demonstrate the various impact of elements of framing theory proving the rationale of applying framing theory to increase the performance in misinformation detection.

Figures (3)

• Figure 1: The Architecture of the Frame Element-based Model
• Figure 2: Measure the performance of removing one of the elements on all four datasets.
• Figure 3: The F1-scores during the training process on all four datasets.