Liars know they are lying: differentiating disinformation from disagreement

The proliferation of misinformation and disinformation poses significant threats to public health, evidence-based policymaking, and democratic processes. This paper directly confronts recent criticisms of misinformation research, which have largely arisen from two sources: a political backlash claiming such research silences conservative voices, and academic critiques questioning the identifiability and prevalence of misinformation. The authors contend that the spread of misinformation, particularly willful disinformation (outright lies), is demonstrably harmful. They aim to show that disinformation is identifiable and differs significantly from honest political debate. The paper concludes by outlining empirically validated, rights-preserving methods to mitigate misinformation and disinformation, without resorting to censorship. The authors use the term 'misinformation' broadly to refer to false information, regardless of intent, while 'disinformation' specifically refers to intentionally false information.

The paper draws upon a vast body of research on misinformation and disinformation, citing numerous studies on the impact of false information on elections, public health, and democratic institutions. It reviews studies examining the asymmetrical spread of misinformation across the political spectrum, with conservatives in the US being disproportionately likely to consume, share, and believe false information. This asymmetry is observed not only in public opinion but also in the behavior of politicians and the content of media outlets. The paper also explores the literature on the cognitive processes underlying belief in misinformation, including the roles of motivated reasoning, emotional appeals, and cognitive biases. Finally, it examines research into the methods used to identify and counter misinformation, including fact-checking, crowdsourcing, and linguistic analysis.

The paper's methodology is primarily a literature review and synthesis of existing research. The authors systematically review studies examining the prevalence, impact, and identifiability of misinformation and disinformation. They analyze case studies, including the "Big Lie" about the 2020 US presidential election, the tobacco industry's disinformation campaign, and ExxonMobil's denial of climate change. They also draw upon research in cognitive psychology, linguistics, and computational social science to assess the methods used to identify and counter misinformation. The paper's analysis relies on a qualitative examination of these case studies and a quantitative review of the results from multiple studies, demonstrating that the findings are consistent and generalizable across various contexts. Natural language processing (NLP) techniques and machine learning models are highlighted as tools for identifying deceptive intent and the linguistic markers of misinformation.

The paper's key findings include: 1. Disinformation is demonstrably harmful and distinguishable from good-faith political disagreement. The "Big Lie" surrounding the 2020 US election serves as a prime example, showing how deliberate disinformation erodes trust in democratic institutions and can incite violence. 2. Misinformation and disinformation are identifiable. While some contested assertions are legitimately debatable, outright lies and propaganda are often distinguishable through various methods. These methods include linguistic analysis, which reveals patterns in language used to convey misinformation, and analysis of internal documents, revealing discrepancies between private knowledge and public pronouncements. Crowdsourcing, despite potential biases, also provides a useful tool for identifying misinformation. 3. There is a significant political asymmetry in the spread and consumption of misinformation, particularly in the US, with conservatives disproportionately consuming and believing false information. This asymmetry is evident in both public opinion and the behavior of political elites and media outlets. 4. Willful disinformation can often be identified by examining discrepancies between public statements and actions in legal settings or internal communications. The authors use examples from various lawsuits against Trump's campaign and Fox News to illustrate this point. 5. A range of empirically validated methods exists for mitigating misinformation and disinformation without resorting to censorship. These include fact-checking, prebunking (inoculation), promoting media literacy, and modifying platform algorithms to reduce the spread of sensationalistic content. 6. The attacks on misinformation research are politically motivated and reminiscent of past attacks on science by powerful interests. The paper argues these attacks should be understood as attempts to silence researchers and undermine efforts to counter disinformation, not as legitimate defenses of free speech.

The findings challenge the notion that misinformation is too difficult to identify or lacks widespread impact to warrant attention. The paper demonstrates that, while some disagreements are legitimately rooted in differing interpretations of evidence, there is a clear distinction between these disagreements and deliberately deceptive disinformation. The significant political asymmetry in misinformation consumption necessitates targeted interventions. The effectiveness of fact-checking, prebunking, and other strategies underscores the importance of continuing research in this area. The paper's discussion highlights the tension between free speech and the need to counter misinformation, arguing that deliberate disinformation, which involves fraud and intentional deception, is not protected by free speech principles. The authors suggest that addressing the problem requires a multi-pronged approach, involving interventions that both improve consumers' ability to discern information and hold purveyors of disinformation accountable.

This paper demonstrates that the spread of misinformation and disinformation poses significant threats to democracy and public well-being. Misinformation and disinformation are identifiable using a variety of methods and can be effectively countered without compromising free speech principles. Future research should explore the long-term societal consequences of misinformation and develop more sophisticated strategies for countering it, particularly in diverse cultural and political contexts.

The paper's primary limitation is its reliance on a literature review. While the authors synthesize a large body of research, further empirical studies may be needed to fully explore some of the complex relationships between misinformation, political polarization, and institutional trust. The specific examples used, such as the "Big Lie" and the activities of Fox News, may not be fully generalizable to all contexts, although the underlying principles of misinformation identification and mitigation are likely to be widely applicable.