Developing Fake News Immunity: Fallacies as Misinformation Triggers During the Pandemic

Abstract

Misinformation constitutes one of the main challenges to counter the infodemic: misleading news, even if not blatantly false, can cause harm especially in crisis scenarios such as the pandemic. Due to the fast proliferation of information across digital media, human fact-checkers struggle to keep up with fake news, while automatic fact-checkers are not able to identify the grey area of misinformation. We, thus, propose to reverse engineer the manipulation of information offering citizens the means to become their own fact-checkers through digital literacy and critical thinking. Through a corpus analysis of fact-checked news about COVID-19, we identify 10 fallacies–arguments which seem valid but are not–that systematically trigger misinformation and offer a systematic procedure to identify them. Next to fallacies, we examine the types of sources associated to (mis-/dis-)information in our dataset as well as the type of claims making up the headlines. The statistical patterns surfaced from these three levels of analysis reveal a misinformation ecosystem where no source type is exempt from flawed arguments with frequent evading the burden of proof and cherry picking behaviors, even when descriptive claims are at stake. In such a scenario, exercising the audience’s critical skills through fallacy and semantic analysis is necessary to guarantee fake news immunity.