Media Contact: Claire Sabin, claire [dot] sabin
gwu [dot] edu (claire[dot]sabin[at]gwu[dot]edu)
WASHINGTON (May 20, 2026) — New research sheds light on how extreme hate content spreads across online platforms and why certain moderation strategies can amplify the problem. The study uses a mathematical model to examine how online communities both form and break apart as harmful content spreads.
The study showed:
- Harmful online content spreads through interconnected clusters of extremist and non-extremist online communities across multiple social media platforms. It spreads like a fire in a forest where the trees move as well as the fire.
- Reducing the number of extremist online communities may initially slow the spread of hate content, but certain moderation efforts can eventually create conditions that allow harmful content to spread again.
- Digital vaccination strategies, i.e. exposing users to factual counter-messaging before they encounter extremist content, may help reduce widespread online spreading.
“Our findings show that online hate ecosystems behave in more complex ways than many current moderation approaches assume,” said Neil Johnson, professor of physics at the George Washington University who led the research. “The network dynamics are complex and require clear interventions and solutions.”
The researchers say the model could eventually help policymakers and technology companies better prepare for how hate content moves online. This could result in more effective interventions before harmful narratives spread.
The paper, “Re-entrant spreading in two-species coalescence-fragmentation with SIR dynamics” was coauthored by Johnson, Chen Xu, professor at Soochow University, Pak Ming Hui, professor at The Chinese University of Hong Kong, and Chenkai Xia, PhD student at GW. It was published in the journal Physical Review E.
If you’re interested in learning more, please contact Claire Sabin at claire [dot] sabingwu [dot] edu (claire[dot]sabin[at]gwu[dot]edu).
-GW-