Exploring the Zombie Internet

Anatomy of Three Deceptive Information Operations on Facebook

Abstract

This study advances the concept of deceptive information operations as an analytical framework that moves beyond traditional false-news taxonomies to capture a broader spectrum of manipulative and harmful communicative activities on social media. Drawing on Starbird et al.'s (2019) work on strategic information operations, Chadwick and Stanyer's (2022) theory of deception and META’s Coordinated Inauthentic behaviour (Gleicher, 2018), we examine three cases discovered through an automated detection workflow that continuously monitors coordinated sharing behaviour on Facebook: a state-aligned pro-Putin propaganda network, a coordinated campaign promoting online gambling, and a network of poorly moderated public groups distributing adult content and fraudulent links. Each case is analysed through three dimensions: the deceptive nature of the content, the exploitation of platform affordances, and the participatory behaviours that sustain and amplify these operations. We used an inductive grounded coding approach applied to account-level metadata and the 500 most-viewed posts per network retrieved via the Meta Content Library. The findings reveal shared mechanisms across all three operations, e.g. coordinated behaviour to exploitat Facebook's algorithmic ranking and simulate organic popularity, and the deliberate mixing of benign and harmful content to obscure intent and evade moderation. These dynamics resonate with the notion of the "Zombie Internet", an information ecosystem in which the boundary between human and automated, toxic behaviour becomes increasingly difficult to maintain. Together, the cases expose structural vulnerabilities in platform governance that allow deceptive operations to persist, scale, and evade detection. The study contributes an empirically grounded conceptualisation of deceptive information operations and identifies critical gaps in platform governance frameworks that focus narrowly on content inauthenticity rather than addressing the structural conditions that enable manipulation at scale.

The findings reveal shared mechanisms across all three operations: coordinated timing of posts, exploitation of Facebook's algorithmic ranking to simulate organic popularity, synthetic engagement, and the deliberate mixing of benign and harmful content to obscure intent and evade moderation. These dynamics resonate with the notion of the "Zombie Internet", an information ecosystem in which the boundary between human and automated, toxic behaviour becomes increasingly difficult to maintain. Together, the cases expose structural vulnerabilities in platform governance that allow deceptive operations to persist, scale, and evade detection. The study contributes an empirically grounded conceptualisation of deceptive information operations and identifies critical gaps in platform governance frameworks that focus narrowly on content inauthenticity rather than addressing the structural conditions that enable manipulation at scale.