Explaining Sources of Uncertainty in Automated Fact-Checking

2025-05-23

AI Summary

In automated fact-checking, model uncertainty—especially under conflicting evidence—lacks interpretability: conventional approaches yield only scalar confidence scores or vague statements, hindering user understanding and trust. To address this, we propose CLUE, a plug-and-play framework requiring no fine-tuning or architectural modification. CLUE employs unsupervised span-level relation identification to explicitly model evidential consistency and conflict, then leverages attention-guided prompting to generate natural-language uncertainty explanations tightly grounded in the underlying reasoning. Its core contribution is the first interpretable modeling of evidential conflict—a primary source of uncertainty in fact-checking. Experiments across three language model families and two benchmark fact-checking datasets demonstrate that CLUE significantly improves explanation faithfulness and decision consistency. Human evaluation further confirms that CLUE’s explanations are more effective, informative, and logically rigorous than baselines.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Understanding sources of a model's uncertainty regarding its predictions is crucial for effective human-AI collaboration. Prior work proposes using numerical uncertainty or hedges ("I'm not sure, but ..."), which do not explain uncertainty that arises from conflicting evidence, leaving users unable to resolve disagreements or rely on the output. We introduce CLUE (Conflict-and-Agreement-aware Language-model Uncertainty Explanations), the first framework to generate natural language explanations of model uncertainty by (i) identifying relationships between spans of text that expose claim-evidence or inter-evidence conflicts and agreements that drive the model's predictive uncertainty in an unsupervised way, and (ii) generating explanations via prompting and attention steering that verbalize these critical interactions. Across three language models and two fact-checking datasets, we show that CLUE produces explanations that are more faithful to the model's uncertainty and more consistent with fact-checking decisions than prompting for uncertainty explanations without span-interaction guidance. Human evaluators judge our explanations to be more helpful, more informative, less redundant, and more logically consistent with the input than this baseline. CLUE requires no fine-tuning or architectural changes, making it plug-and-play for any white-box language model. By explicitly linking uncertainty to evidence conflicts, it offers practical support for fact-checking and generalises readily to other tasks that require reasoning over complex information.

Problem

• Explaining model uncertainty in automated fact-checking systems
• Identifying conflicts and agreements in evidence causing uncertainty
• Generating natural language explanations for model uncertainty

Innovation

• Unsupervised identification of claim-evidence conflicts
• Generating explanations via prompting and attention steering
• Plug-and-play for any white-box language model