How We Refute Claims: Automatic Fact-Checking through Flaw Identification and Explanation

TL;DR

The paper addresses scalable automatic fact-checking by introducing flaw-oriented reasoning through aspect generation and a seven-flaw taxonomy, backed by the FlawCheck dataset and the RefuteClaim framework. It demonstrates that jointly modeling aspects, flaws, and evidence improves both justification generation and veracity classification, with notable gains for false and partly false claims, while Unproven remains challenging. The approach aims for interpretable, evidence-grounded outputs suitable for deployment and human-in-the-loop workflows, advancing beyond traditional veracity labeling toward explainable misinformation analysis.

Abstract

Automated fact-checking is a crucial task in the governance of internet content. Although various studies utilize advanced models to tackle this issue, a significant gap persists in addressing complex real-world rumors and deceptive claims. To address this challenge, this paper explores the novel task of flaw-oriented fact-checking, including aspect generation and flaw identification. We also introduce RefuteClaim, a new framework designed specifically for this task. Given the absence of an existing dataset, we present FlawCheck, a dataset created by extracting and transforming insights from expert reviews into relevant aspects and identified flaws. The experimental results underscore the efficacy of RefuteClaim, particularly in classifying and elucidating false claims.