Learning from Fact-checkers: Analysis and Generation of Fact-checking Language

TL;DR

This paper introduces a deep learning framework that generates fact-checking responses to online misinformation, leveraging linguistic analysis of fact-checking tweets to enhance engagement and combat fake news effectively.

Contribution

It presents a novel application of text generation for fact-checking, analyzing linguistic features and developing a model that outperforms existing methods in generating fact-checking responses.

Findings

Fact-checkers tend to refute misinformation and use formal language.

The proposed framework achieves up to 30% improvement over competing models.

Generated responses are qualitatively superior to existing models.

Abstract

In fighting against fake news, many fact-checking systems comprised of human-based fact-checking sites (e.g., snopes.com and politifact.com) and automatic detection systems have been developed in recent years. However, online users still keep sharing fake news even when it has been debunked. It means that early fake news detection may be insufficient and we need another complementary approach to mitigate the spread of misinformation. In this paper, we introduce a novel application of text generation for combating fake news. In particular, we (1) leverage online users named \emph{fact-checkers}, who cite fact-checking sites as credible evidences to fact-check information in public discourse; (2) analyze linguistic characteristics of fact-checking tweets; and (3) propose and build a deep learning framework to generate responses with fact-checking intention to increase the fact-checkers' engagement in fact-checking activities. Our analysis reveals that the fact-checkers tend to refute misinformation and use formal language (e.g. few swear words and Internet slangs). Our framework successfully generates relevant responses, and outperforms competing models by achieving up to 30\% improvements. Our qualitative study also confirms that the superiority of our generated responses compared with responses generated from the existing models.