Assessing the risk of “infodemics” in response to COVID-19 epidemics

Our society is built on a complex web of interdependen- cies whose effects become manifest during extraordinary events, with shocks in one system propagating to the others to an exceptional extent. The recent explosion of publicly shared, decentralized information production that character- izes digital societies and in particular social media activity provides an exceptional laboratory for the observation and the study of these complex social dynamics, and potentially functions as a laboratory to understand, test and validate possible solutions to large-scale crises. Global pandemics are certainly an instance of such crises, and the current out- break of COVID-19 may therefore be thought of as a natu- ral experiment to observe social responses to a major threat that may potentially escalate to catastrophic levels, and has already managed to seriously affect levels of economic ac- tivity, and radically alter human social behaviors across the globe.
In this study [1], we analyzed more than 100 millions Twitter messages posted worldwide in 64 languages during the epidemic emergency due to SARS-CoV-2 and classified the reliability of news diffused to show that information dy- namics tailored to alter individuals’ perceptions and, con- sequently, their behavioral response, is able to drive collec- tive attention towards false or inflammatory content, a phe- nomenon named infodemics, sharing similarities with more traditional epidemics and spreading phenomena.
For each message, we identify the presence of links point- ing to external websites: for each link we verify if it comes from a trustworthy source or not by aggregating information by several fact-checking databases. The number of followers of a single user is then used defines the exposure in terms of potential visualizations at first-order approximation. Expo- sure and reliability are useful descriptors that, however, do not capture alone the risk of infodemics. For this reason we have developed an Infodemic Risk Index (IRI) which quan- tifies the rate at which a generic user is exposed to unreliable news produced by a specific class of users (partial IRI) or by any class of users (IRI).
Tracking the evolution in time of the IRI across the globe, we find that, contrary to what it could be expected in prin- ciple, on the verge of a threatening global pandemic emer- gency, human communication activity is to a significant ex- tent characterized by the intentional production of informa- tional noise and even of misleading or false information. This generates waves of unreliable and low-quality informa- tion with potentially very dangerous impacts on the social capacity to respond adaptively at all scales by rapidly adopt- ing those norms and behaviors that may effectively contain the propagation of the epidemics.
Fortunately, we also find that the escalation of the epi- demics leads people to progressively pay attention to more reliable sources thus potentially limiting the impact of the
infodemics, but the actual speed of adjustment may make a major difference in determining the social outcome, and in particular between a controlled epidemics and a global pan- demics.
This casts new light on the social mechanics of the infodemics-epidemics interaction, and may be of help to policy makers to design a more integrated strategic ap- proach, by suitably embedding communication and infor- mation management into a comprehensive, extended public health perspective.