Abstract
Introduction
The spread of mis- and disinformation is an increasing concern for democratic societies around the globe (Lazer et al., 2019). In modern high-choice media environments, exposure to misinformation can be harmful and can have negative consequences for democratic governance as well as trust in news media and journalism more broadly (e.g., Bennett & Livingston, 2018; Chesney & Citron, 2018; Nisbet et al., 2021; Ognyanova et al., 2020; Vaccari & Chadwick, 2020). Some scholars have even argued that we are entering a post-truth era with an alternative epistemology and thus an alternative reality, in which, for instance, former president Obama was not born in the United States and global warming is simply a Chinese hoax (Lewandowsky et al., 2017) rather than largely undisputed scientific phenomenon (Cook et al., 2016).
These ongoing “debates” highlight the importance of several distinct but related constructs. The first is the concept of “fake news,” which refers to false or misleading information (Lazer et al., 2019; see Tsfati et al., 2020). Egelhofer and Lecheler (2019) proposed that fake news can be conceptualized as a two-dimensional phenomenon differentiating (a) fake news genre or “the deliberate creation of pseudojournalistic disinformation” (p. 97) from (b) the fake news label used (e.g., by politicians like Donald Trump) to delegitimize news media. Second, there is a distinction between disinformation and misinformation, whereas disinformation is a subset of misinformation. Disinformation is spread intentionally by various actors who know that the information is false. In contrast, misinformation is spread by actors who mistakenly believe the information to be factually correct when it is not (Vaccari & Chadwick, 2020; Vraga & Bode, 2020).
Previous research has primarily focused on textual forms of misinformation, while visual and multimodal forms (e.g., news images, memes, and videos) of misinformation have received much less attention. This is surprising because visual information may affect how media consumers select and process information (Zillmann et al., 2001; see also Garcia & Stark, 1991; Sargent, 2007). Furthermore, visual information can affect news consumers’ emotional reactions (e.g., Iyer et al., 2014), attitudes (Matthes et al., 2021; Powell et al., 2015; von Sikorski, 2021; von Sikorski & Ludwig, 2018), and behavioral responses (Powell et al., 2015) independently of textual communication (von Sikorski & Knoll, 2019). This tendency is likely due to visual information coming “with an implicit guarantee of being closer to the truth than other forms of communication” (Messaris & Abraham, 2001, p. 217). Thus, visual mis- and disinformation may be particularly persuading (see Messaris, 1997) and could have damaging effects for democratic governance. Visual information can be manipulated or taken out-of-context and can be (mis)used as a credible type of “proof” (e.g., “deepfakes” of politicians). Emerging research has shown that new multimodal forms of misinformation are disseminated quickly and seamlessly via social media and can have considerable negative effects on political attitudes and decision-making (Hameleers et al., 2020; Vaccari & Chadwick, 2020).
For instance, focusing on the period leading up to the 2019 Indian national elections, Garimella and Eckles (2020) showed that 13% of all images shared on WhatsApp public groups in India qualified as visual misinformation (for visuals in COVID-19 misinformation, see Brennen et al., 2020). In a bottom-up approach, ordinary users can “produce” and spread mis- and disinformation via social media on their own by manipulating photographs or by using simple editing techniques to manipulate original video material (slowing down a sound-track, de-/re-contextualizing visual information, etc.). However, mis- and disinformation can also be spread top-down. For instance, political actors can disseminate mis- and disinformation to their followers via social media and thus quickly reach large audiences, bypassing mainstream media outlets, using both nonsophisticated forms of mis- and disinformation (e.g., out-of-context visual information) and sophisticated manipulation techniques like “deepfake” videos based on artificial intelligence and machine learning procedures (Vaccari & Chadwick, 2020; for an example, also see Christopher, 2020). Although, bottom-up and top-down dissemination processes can generally be differentiated, multimodal misinformation may further spread through social media networks in complex ways, enabling political actors to circulate and further disseminate misinformation created by ordinary citizens or political groups (for an example, see Harwell, 2019). Yet, social media platforms are not the only media sources that influence whether and how multimodal mis- and disinformation spreads (Allcott et al., 2019; Donovan, 2021; Guess, Nyhan, et al., 2020). Tsfati and colleagues (2020) emphasized the importance of mainstream media in the spread of mis- and disinformation, as citizens regularly learn about political disinformation campaigns via mainstream media coverage.
Visual political mis- and disinformation is still not well understood by the scholarly community as scientific research about this phenomenon is still in its infancy—leaving many questions unanswered. For instance, how can visual misinformation be effectively debunked (Hameleers et al., 2020; Young et al., 2018)? How can backfire effects and continued influence effects of misinformation be best prevented (e.g., Lewandowsky et al., 2020; Nyhan, 2021; Stubenvoll & Matthes, 2021)? Are there ways to inoculate individuals against (visual) misinformation (Basol et al., 2021; Compton et al., 2021)?
The aim of the invited forum is to find answers to some of these questions and to bring together leading researchers from the fields of political communication, visual communication, psychology, and data science to provide a comprehensive overview of the state of research, noting key challenges and identifying avenues for future research. The forum brings together expert scholars focusing on key domains of visual political mis- and disinformation. Viorela Dan (University of Munich) focuses on different types of disinformation videos and challenges for journalism and democracy. Britt Paris (Rutgers University) and Joan Donovan (Harvard University) examine online platform functionality and the fight against audiovisual disinformation. Michael Hameleers (University of Amsterdam) points out the effects of multimodal disinformation, and how it can be potentially debunked and efficiently corrected. Based on inoculation theory, Jon Roozenbeek and Sander van der Linden (University of Cambridge) examine an innovative way of prebunking or “vaccinating citizens against visual disinformation” before media users are exposed to multimodal falsehoods. Finally, I will point out “next steps” and future avenues for research on multimodal misinformation. In all, these contributions offer important insights and clarify why we should continue to research and expand our knowledge of multimodal mis- and disinformation in the future.