Malgré l'attention accordée à l'enjeu de la mésinformation au cours des dernières années, peu d’études ont examiné l'appui des citoyens pour les mesures visant à y faire face. À l'aide de données récoltées lors des élections québécoises de 2022 et de modèles par blocs récursifs, cet article montre que l'appui aux interventions contre la mésinformation est élevé en général, mais que les individus ayant une idéologie de droite, appuyant le Parti conservateur du Québec et n'ayant pas confiance dans les médias et les scientifiques sont plus susceptibles de s'y opposer. Ceux qui ne sont pas préoccupés par l'enjeu, priorisent la protection de la liberté d'expression ou adhèrent aux fausses informations y sont aussi moins favorables. Les résultats suggèrent que dépolitiser l'enjeu de la mésinformation et travailler à renforcer la confiance envers les institutions pourraient augmenter la légitimité perçue et l'efficacité de notre réponse face à la mésinformation.

Armed conflict presents a multitude of risks to civilians, prisoners of war and others caught in the middle of hostilities. Harmful information spreading on social media compounds such risks in a variety of tangible ways, from potentially influencing acts that cause physical harm to undermining a person's financial stability, contributing to psychological distress, spurring social ostracization and eroding societal trust in evidentiary standards, among many others. Despite this span of risks, no typology exists that maps the full range of such harms. This article attempts to fill this gap, proposing a typology of harms related to the spread of harmful information on social media platforms experienced by persons affected by armed conflict. Developed using real-world examples, it divides potential harm into five categories: harms to life and physical well-being, harms to economic or financial well-being, harms to psychological well-being, harms to social inclusion or cultural well-being, and society-wide harms. After detailing each component of the typology, the article concludes by laying out several implications, including the need to view harmful information as a protection risk, the importance of a conflict-specific approach to harmful information, the relevance of several provisions under international law, and the possible long-term consequences for societies from harmful information.

The information used for this typology is based entirely on open-source reporting covering acts that occurred during armed conflict and that were seemingly related to identified harmful information on social media platforms or messaging applications. The authors did not verify any reported incidents or information beyond what was included in cited sources. Throughout the article, sources have been redacted from citations where there is a risk of reprinting harmful information or further propagating it, and where redaction was necessary to avoid the impression that the authors were attributing acts to particular groups or actors.

This review article explores the role of land-grant Extension amidst an escalating epistemic crisis, where misinformation and the contestation of knowledge severely impact public trust and policymaking. We delve into the historical mission of land-grant institutions to democratize education and extend knowledge through Cooperative Extension Services, highlighting their unique position to address contemporary challenges of information disorder and declining public confidence in higher education. Land-grant universities can reaffirm their relevance and leadership in disseminating reliable information by reasserting their foundational principles of unbiased, objective scholarship and deep engagement with diverse stakeholders. This reaffirmation comes at a critical time when societal trust in science and academia is waning, necessitating a recommitment to community engagement and producing knowledge for the public good. The article underscores the necessity for these institutions to adapt to the changing information landscape by fostering stakeholder-engaged scholarship and enhancing accessibility, thus reinforcing their vital role in upholding the integrity of public discourse and policy.

Text classification methods have been widely investigated as a way to detect content of low credibility: fake news, social media bots, propaganda, etc. Quite accurate models (likely based on deep neural networks) help in moderating public electronic platforms and often cause content creators to face rejection of their submissions or removal of already published texts. Having the incentive to evade further detection, content creators try to come up with a slightly modified version of the text (known as an attack with an adversarial example) that exploit the weaknesses of classifiers and result in a different output. Here we systematically test the robustness of common text classifiers against available attacking techniques and discover that, indeed, meaning-preserving changes in input text can mislead the models. The approaches we test focus on finding vulnerable spans in text and replacing individual characters or words, taking into account the similarity between the original and replacement content. We also introduce BODEGA: a benchmark for testing both victim models and attack methods on four misinformation detection tasks in an evaluation framework designed to simulate real use cases of content moderation. The attacked tasks include (1) fact checking and detection of (2) hyperpartisan news, (3) propaganda, and (4) rumours. Our experimental results show that modern large language models are often more vulnerable to attacks than previous, smaller solutions, e.g. attacks on GEMMA being up to 27% more successful than those on BERT. Finally, we manually analyse a subset adversarial examples and check what kinds of modifications are used in successful attacks.

We introduce a dynamic dataset of all communications by state election officials (EOs) on social media during the 2022 election cycle and develop metrics to assess the effectiveness of trust-building strategies on voter confidence. We employ quantitative manual content analysis of 10,000 organic posts from 118 state EOs’ accounts on Facebook, Instagram, and Twitter between September 10 and November 30, 2022, and code for the presence of variables that measure EOs’ efforts to combat misinformation and build trusted networks of communications. The measures we present here address two questions: (1) How much coordination was there among states in terms of incorporating the #TrustedInfo2022 campaign, promoted by the National Association of Secretaries of State, in their social media communications, and (2) How much of states’ social media communications explicitly signaled that EOs are trusted sources of information? We demonstrate the applicability of our data on research that evaluates the impact of trust-building campaigns on voter confidence in elections, which is grounded on theories of deliberative democracy and democratic listening.

To what extent can the harms of misinformation be mitigated by relying on nudges? Prior research has demonstrated that non-intrusive ‘accuracy nudges’ can reduce the sharing of misinformation. We investigate an alternative approach. Rather than subtly reminding people about accuracy, our intervention, indebted to research on the bystander effect, explicitly appeals to individuals' capacity to help solve the misinformation challenge. Our results are mixed. On the one hand, our intervention reduces the willingness to share and believe in misinformation fact-checked as false. On the other hand, it also reduces participants' willingness to share information that has been fact-checked as true and innocuous, as well as non-fact-checked information. Experiment 1 offers proof of concept; Experiment 2 tests our intervention with a more realistic mix of true and false social media posts; Experiment 3 tests our interventions alongside an accuracy nudge. The effectiveness of our intervention at reducing willingness to share misinformation remains consistent across experiments; meta-analysis reveals that our treatment reduced willingness to share false content across experiments by 20% of a scale point on a six-point scale. We do not observe the accuracy nudge reducing willingness to share false content. Taken together, these results highlight the advantages and disadvantages of accuracy nudges and our more confrontational approach.

The field of misinformation studies has experienced a boom of scholarship in recent years. Buoyed by the emergence of information operations surrounding the 2016 election and the rise of so-called “fake news,” researchers hailing from fields ranging from philosophy to computer science have taken up the challenge of detecting, analyzing, and theorizing false and misleading information online. In an attempt to understand the spread of misinformation online, researchers have adapted concepts from different disciplines. Concepts from epidemiology, for example, have opened doors to thinking about spread, contagion, and resistance. The life sciences offer concepts and theories to further extend what we know about how misinformation adapts; by viewing information as an organism within a complex ecosystem, we can better understand why some narratives succeed and others fail. Collaborations between misinformation researchers and life scientists to develop responsible adaptations of fitness models can bolster misinformation research.

In a “mixed bag” 2023-2024 session, the U.S. Supreme Court issued a series of decisions both favorable and antithetical to public health and safety. Taking on tough constitutional issues implicating gun control, misinformation, and homelessness, the Court also avoided substantive reviews in favor of procedural dismissals in key cases involving reproductive rights and government censorship.

Does nationalism increase beliefs in conspiracy theories that frame minorities as subversives? From China to Russia to India, analysts and public commentators increasingly assume that nationalism fuels belief in false or unverified information. Yet existing scholarly work has neither theoretically nor empirically examined this link. Using a survey experiment conducted among 2,373 individuals and 6 focus groups with 6–8 participants each, for a total of 50 individuals, we study the impact of nationalist sentiment on belief in conspiracy theories related to ethnic minority groups in Pakistan. We find that nationalist primes – even those intended to emphasize the integration of diverse groups into one superordinate national identity – increase belief in statements about domestic minorities collaborating with hostile foreign powers. Subgroup analysis and focus groups suggest that nationalism potentially increases the likelihood that one views rights-seeking minorities as undermining the pursuit of national status.

During health crises, misinformation may spread rapidly on social media, leading to hesitancy towards health authorities. The COVID-19 pandemic prompted significant research on how communication from health authorities can effectively facilitate compliance with health-related behavioral advice such as distancing and vaccination. Far fewer studies have assessed whether and how public health communication can help citizens avoid the harmful consequences of exposure to COVID-19 misinformation, including passing it on to others. In two experiments in Denmark during the pandemic, the effectiveness of a 3-minute and a 15-second intervention from the Danish Health Authorities on social media was assessed, along with an accuracy nudge. The findings showed that the 3-minute intervention providing competences through concrete and actionable advice decreased sharing of COVID-19-related misinformation and boosted their sense of self-efficacy. These findings suggest that authorities can effectively invest in building citizens’ competences in order to mitigate the spread of misinformation on social media.