Nonlinear spreading behavior across multi-platform social media universe

Area of Science:

• Computational Social Science
• Network Science
• Mathematical Modeling

Background:

• Online communities dynamically interconnect across multiple social media platforms.
• Understanding the viral spread of harmful content (misinformation, hate speech) is a critical societal challenge.
• Existing models often do not fully capture cross-platform community dynamics.

Purpose of the Study:

• To develop a non-linear dynamical model for viral spreading of harmful content across interconnected online communities.
• To identify analytic conditions for the onset of system-wide outbreaks.
• To inform policy interventions against the spread of harmful online content.

Main Methods:

• Development of a non-linear dynamical model.
• Application of mean-field theory (Effective Medium Theory).
• Comparison with detailed numerical simulations.

Main Results:

• The model accurately predicts viral spreading dynamics.
• An analytic condition for outbreak onset was derived.
• System-wide spreading is predicted when inter-community linking rates are high and link loss rates are low, irrespective of infection vs. recovery rates.

Conclusions:

• High rates of inter-community linking and low rates of link loss can lead to widespread harmful content outbreaks.
• Effective Medium Theory provides a robust framework for analyzing cross-platform content dynamics.
• Policymakers must consider multi-community dynamics when designing strategies to combat online misinformation and hate speech.