Coordinated Inauthentic Behavior and Information Spreading on Twitter

Matteo Cinelli, Stefano Cresci, Walter Quattrociocchi, Maurizio Tesconi, Paola Zola · 2025

Read Coordinated Inauthentic Behavior and Information Spreading on Twitter by Matteo Cinelli, Stefano Cresci, Walter Quattrociocchi, Maurizio Tesconi, Paola Zola (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-15720.txt.

Introduction

Matteo Cinelli, Stefano Cresci, Walter Quattrociocchi, Maurizio Tesconi, and Paola Zola study fake teamwork on Twitter in a 2025 paper. Some accounts act in sync to push posts further than normal users can. The team calls this coordinated inauthentic behavior. They ask how much it changes the spread of posts.

They tracked large chains of retweets and marked the synced accounts. Synced accounts sat near the start of chains and spread posts faster. Chains with more synced accounts grew larger and taller than chance would suggest. Two new measures track how well synced accounts pull in normal users. Past a share of about half, more synced accounts added no extra reach. The work helps explain how fake teamwork shapes what people see.

Catalog record

CREATORMatteo Cinelli, Stefano Cresci, Walter Quattrociocchi, Maurizio Tesconi, Paola Zola
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-15720

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Coordinated Inauthentic Behavior and Information Spreading on Twitter Matteo Cinelli, Stefano Cresci, Walter Quattrociocchi, Maurizio Tesconi, Paola Zola Abstract We explore the effects of coordinated users (i.e., users characterized by an unexpected, suspicious, or exceptional similarity) in information spreading on Twitter by quantifying the efficacy of their tactics in deceiving feed algorithms to maximize information outreach. In particular, we investigate the behavior of coordinated accounts within a large set of retweet-based information cascades identifying key differences between coordinated and non-coordinated accounts in terms of position within the cascade, action delay and outreach. On average, coordinated accounts occupy higher positions of the information cascade (i.e., closer to the root), spread messages faster and involve a slightly higher number of users. When considering cascade metrics such as size, number of edges and height, we observe clear differences among information cascades that are associated to a systematically larger proportion of coordinated accounts, as confirmed by comparisons with statistical null models. To further characterize the activity of coordinated accounts we introduce two new measures capturing their infectivity within the information cascade (i.e., their ability to involve other users) and their interaction with non-coordinated accounts. Finally, we find that the interaction pattern between the two classes of users follows a saturation-like process. A larger-scale targeting of non-coordinated users does not require a larger amount of coordinated accounts after a threshold value ∼ 50 % \sim 50\% , after which involving more coordinated accounts within a cascade yields a null marginal effect. Our results contribute to shed light on the role of coordinated accounts and their effect on information diffusion. 1 Introduction Social media play a pivotal role in the evolution of public debates and in shaping public opinion. Platforms that were originally designed for entertainment are nowadays one of the main gateways to information sources and the principal environment in which public opinions get shaped. Users can access and share an unprecedented amount of information, and feed algorithms affect the selection process. Such a constant information overload determines a new tendency of news consumption [ 1 , 2 , 3 ] , and generates opportunities for civic engagement and public interest [ 4 ] , also eliciting the so-called “democratizing effect”. However, social media are also affected by a number of ailments, including toxicity, harassment and hateful speech [ 5 , 6 ] , polarization and radicalization [ 7 ] , propaganda and misleading information [ 8 ] . Regarding the latter issue, many studies recently investigated the spread of fake information and propaganda [ 9 , 10 ] and the role of disinformation on social media [ 11 , 12 , 13 ] . Within this context, scholars also investigated the role of automated accounts in spreading both reliable and questionable information [ 14 , 15 ] . This literature highlights that whether initial social bots could be accurately identified by state-of-the-art algorithms, recent manipulation campaigns carried out with modern bots and with human-operated accounts (i.e., trolls) are more sophisticated and thus harder to detect [ 16 , 17 ] . Going beyond early literature that focused on the nature (automated or otherwise) of the accounts as a way to distinguish between malicious and legitimate actions, more recent research moved the focus on the actions themselves and on the behaviors of the accounts. In a recent report, Facebook introduced the concept of false amplifiers that are characterized by “coordinated activity by inauthentic accounts with the intent of manipulating political discussions” [ 18 ] . The Facebook report introduces two new concepts associated with online information manipulation: coordination and inauthenticity . In literature, coordination is defined as an unexpected, suspicious, or exceptional similarity between a number of users [ 19 ] . In practice, it is often measured as the number of times two accounts behaved similarly, such as when they repeatedly retweet the same posts [ 20 , 19 ] . Coordination among a certain number of users is considered to be a fundamental ingredient in successful manipulations, since it is necessary for the manipulation to obtain a significant outreach. On the contrary, the importance of inauthenticity -- that is, whether online users misrepresent themselves -- is still debated 1 1 1 https://slate.com/technology/2020/07/coordinated-inauthentic-behavior-facebook-twitter.html . To this end, harmfulness , rather than inauthenticity, appears to be more relevant for the analysis of online manipulations 2 2 2 https://medium.com/1st-draft/how-to-improve-our-analysis-of-coordinated-inauthentic-behavior-a4ec62ce9bff . Interestingly, coordination is an orthogonal concept with respect to the long-studied automation 3 3 3 https://m.facebook.com/communitystandards/inauthentic_behavior/ . In fact, recent results demonstrated limited correlations [ 8 ] , or no correlations at all [ 19 ] , between coordinated and automated accounts. In other words, coordination can be achieved without resorting to automation (i.e., social bots). Coordinated harassment by trolls and the supporting activity of online fandoms are two examples – harmful and harmless, respectively – of coordinated activities by human-operated accounts. For these reasons, the study of coordinated accounts and of their role in online information spreading can complement established studies on online information manipulation and can inform the development of new and effective decision support systems against these manipulations. In this paper, we focus on the role of coordinated accounts in information spreading online – a crucial, yet understudied, issue. In particular, we address the efficacy of their coordinated behavior in deceiving feed algorithms to maximize information outreach. We perform an analysis of a large dataset of 1.4 million tweets about the 2019 UK political elections. We find that coordinated and non-coordinated accounts significantly differ in terms of their retweeting activity and their positioning within information cascades. Additionally, their involvement in the cascades is much higher in empirical data that in randomized models. Finally, by analyzing the interactions between the two types of accounts, we uncover a saturation-like process due to which the maximal interaction between coordinated and non-coordinated accounts amounts to \sim 50\% of the cascade links, regardless of the penetration of coordinated accounts in the whole cascade. The paper is structured as follows: Section 2 outlines our research objectives, the main findings and their significance. Section 3 presents related works on coordinated behavior and information spreading. Section 4 illustrates the dataset used for conducting the analysis and the employed measures and methods. Section 5 details the results of the analysis. Section 6 discusses the implications of our research. Finally, Section 7 reports the conclusions. 2 Summary of research objectives, findings and significance 2.1 Research objectives and findings Our study on the differences between the interaction patterns of coordinated and non-coordinated accounts focuses on retweet dynamics, the primary way to spread information on Twitter [ 21 ] . Under this perspective, our overarching goal is that of investigating the role, impact and tactics of coordinated accounts in information spreading on Twitter . In particular, we currently have limited knowledge of the peculiar traits, actions and tactics of coordinated accounts, except for their similarity in posting dynamics. Our work contributes to filling this gap by adding new insights over the limited existing knowledge. We reach our overarching goal by achieving the following three research objectives: RO1: Describe the role of coordinated accounts in information spreading. In fulfillment of RO1, we identify three measures that significantly distinguish the activity of coordinated accounts from that of

Read the complete text