A George Washington University research team created a novel formula that uses a new realm of physics to demonstrate how, why and when hate speech spreads throughout social media.
The first-of-its-kind study, which was published in the journal “Physical Review Letters,” uses a new scientific equation to register the shockwave effect created by bigoted content across online communities. It predicts how groups form, grow and scatter when threatened by moderators—only to re-emerge across platforms.
“Why is it that this problem of online harms—with all the money and effort that’s thrown at eliminating it—still persists?” noted Neil Johnson, professor of physics at GW and co-author of the study. “Because to fight it, you have to find it and understand it. That’s what this equation is able to do.”
Calling the study the “culmination” of nearly a decade of research into online hate activity, Johnson said the new equation is designed as a tool for moderators to predict the shockwave-like spread of hateful content and develop methods to delay, divert and prevent it from spiraling out of control. The formula applies to lightly moderated websites such as 4Chan as well as highly regulated social platforms like Facebook. He also envisioned lawmakers, scientists and other groups using data from the equation to mitigate the influence of extreme hate communities.
“This study presents the missing science of how harms thrive online and, hence, how they can be overcome,” Johnson said. “This missing science is a new form of shockwave physics.”
The equation address the frustrations of groups attempting to curtail the spread of online hate activity. It builds on physics principles to understand how hate speech ripples through online communities, following a different pattern than non-hateful content. While hate communities often seem to spring up unexpectedly, Johnson said they actually quickly acquire members with extreme content. Just as quickly, however, they attract moderator attention and scatter when identified. But they are often able to rapidly reform, either with the same members or adding new ones.
The patterns follow the competing physics forces of fusion and fission—accounting for the spontaneous emergence of hate communities through the absorption of other communities and interested individuals (fusion) and the disciplinary measures moderators take against users and groups that violate a platform’s rules (fission).
“The online hate communities are like a living thing that is evolving,” Johnson said. “If you don’t know how it’s evolving, you are always going to be playing catch up.”
The novel theory could also be applied beyond social media platforms and online message boards, potentially powering moderation strategies on blockchain platforms, generative AI and the metaverse, Johnson explained.
“The breakthrough of this new science is that it describes this huge problem and establishes a basis for advising on what kind of mitigations might work to address it,” he said.
The GW team includes research scientists Pedro D. Manrique and Minzhang Zheng from the Columbian College of Arts and Sciences Physics Department and Ph.D. students Frank Yingjie Huo, Sara El Oud and Lucia Illari. The study was supported by funding from the U.S. Air Force Office of Scientific Research.