The spread of low-credibility content by social bots

Abstract

The massive spread of digital misinformation has been identified as a major threat to democracies. Communication, cognitive, social, and computer scientists are studying the complex causes for the viral diffusion of misinformation, while online platforms are beginning to deploy countermeasures. Little systematic, data-based evidence has been published to guide these efforts. Here we analyze 14 million messages spreading 400 thousand articles on Twitter during ten months in 2016 and 2017. We find evidence that social bots played a disproportionate role in spreading articles from low-credibility sources. Bots amplify such content in the early spreading moments, before an article goes viral. They also target users with many followers through replies and mentions. Humans are vulnerable to this manipulation, resharing content posted by bots. Successful low-credibility sources are heavily supported by social bots. These results suggest that curbing social bots may be an effective strategy for mitigating the spread of online misinformation.

Arming the public with artificial intelligence to counter social bots

“…The spread of fake news online is another area in which the effect of bots is believed to be relevant (Bessi et al, ; Lazer et al, ). A study based on 14 million tweets posted during and after the 2016 U.S. presidential election revealed that bots played a key role in the spread of low‐credibility content (Shao et al, ,b). The study uncovered strategies by which social bots target influential accounts and amplify misinformation in the early stages of spreading, before it becomes viral.…”

“…For example, maximum‐likelihood estimation is employed by Varol, Ferrara, Davis, et al () to find the appropriate cutoff for binary classification: the best score above which an account is classified as a bot and below which it is classified as human. This threshold approach is useful for scientists counting bots in a collection (Gorwa, ; Haustein et al, ; Shao, Ciampaglia, et al, ; Stella et al, ; Suárez‐Serrato, Roberts, Davis, & Menczer, ), but it is not intuitive for humans evaluating a single account.…”

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

“…In this study, we show that information dynamics tailored to alter individuals' perceptions and, consequently, their behavioral response, is able to drive collective attention 12 towards false 13,14 or inflammatory 15 content, a phenomenon named infodemics [16][17][18][19] , sharing similarities with more traditional epidemics and spreading phenomena [20][21][22] . Contrary to what it could be expected in principle, what this natural experiment reveals is that, on the verge of a threatening global pandemic emergency due to SARS-CoV-2 [23][24][25] , human communication activity is to a significant extent characterized by the intentional production of informational noise and even of misleading or false information 26 .…”

Fake news detection using deep learning models: A novel approach

“…Their prime targets are those users who tend to show some influence on others (particularly their social media followers). It was also noticed that the bots that were used to spread false information to the target audience were usually active in the early phases of the fake news propagation and as such, they attracted humans of similar thinking who reposted the same news in their feed . It has similarly been found that social bots often populate the social space to cause harm and deceive social media users.…”