Mass Manipulation in Simulated Social Networks: Dominating vs. Diversifying Attention

TL;DR

It is found that under acquaintance-based topics, a central MIA advances its propaganda effectively, causing volatile and polarized opinions through repeated exposure and echo chambers, and under randomized topics, this leverage disappears: the MIA's influence collapses across all network structures, and opinions become stable and broadly aligned with reality.

Abstract

Modern information environments, especially social media, are highly complex systems that exceed individual processing capacities such as humans' limited attention. This environment/cognition mismatch can increase susceptibility to misinformation, which various actors exploit for anti-social (including anti-democratic or anti-science) aims. This raises the question of how to feasibly sustain societal resilience against misinformation, though the challenge is to find strategies that respect individuals' cognitive limitations. We investigate whether a simple behavioral rule - topic diversification - can enhance collective performance and mitigate vulnerability. In an agent-based model that includes a deceptive mass-influencing agent (MIA), we compare two attention-distribution strategies: (A) acquaintance-based topic selection, where agents return to familiar content, and (B) randomized topics, which diversify attention. We also track dynamics across different network structures. We find that under acquaintance-based topics, a central MIA advances its propaganda effectively, causing volatile and polarized opinions through repeated exposure and echo chambers. Under randomized topics, this leverage disappears: the MIA's influence collapses across all network structures, and opinions become stable and broadly aligned with reality. These results, while deriving from simple simulations, align with realistic theories of bounded rationality and collective cognition, further suggesting a cognitively feasible, easy-to-monitor and robust strategy: distribute attention to combat misinformation.

Figures (15)

• Figure 1: Communication structure in the RGS with initiator $a$, receiver(s) $b$ and topic $c$, as well as exemplary statements about $c$'s reputation. Left: one-to-one interaction, Right: one-to-many interaction.
• Figure 2: Exemplary networks of single simulations with different parameter settings as given in Table \ref{['tab:network_structures']}. The first row shows idealized network types and the second row the realizations thereof as they emerge in the RGS with only ordinary agents. In the bottom row, we replaced ordinary agent 0 with a MIA. Colors represent agents' reputations and are shown as continuous transitions between green (high), yellow (medium) and dark brown (low). The underlying truth---agents' intrinsic honesties---can be read from their numbers (0: $0\%$ honest to 49: $100\%$ honest). Line intensities (and distances between agents as far as possible within a 2D representation) indicate how often two agents talked to each other. Node sizes indicate the degree of an agent (equation \ref{['eq:degree']}) in log scale.
• Figure 3: Distribution of agent 0's reputations as seen by all others at the end of simulation runs, for each network type. Agent 0 can either be an ordinary agent (solid line) or a MIA (filled histogram).
• Figure 4: Each point shows a MIA's reputation and its degree at the end of a single simulation run, with panel insets showing the corresponding correlation across all runs within a given network type.
• Figure 5: Time evolution of the MIA's degree, for each of the three network types, with lines connecting data from a single simulation run. Line color indicates a MIA's reputation at any point in time, ranging from dark brown ($0\%$) over yellow ($50\%$) to green ($100\%$).