A.I. scans social media to predict when protests will turn violent

With all the hostility on social media, it may sound oxymoronic to say that social media platforms could curb violence.

But studies have shown it could help fight wildlife poaching, and now new research from a new study out of the University of Southern California (USC) suggests social media may help avoid violence between humans as well.

USC psychologists and computer scientists created an artificial intelligence (A.I.) algorithm that scanned posts and correlated their content with impending violence at protests. Such a tool could be used to better prepare for demonstrations that are prone to escalation.

By analyzing posts on Twitter, the algorithm was able to pick up on language that signaled hostilities. The research also revealed that moral topics (issues that users regarded as blatantly right or wrong) were most likely to ignite violence.

“Our findings suggest that people are more likely to condone violent protest of an issue when they both see it at as a moral issue and believe others share this position, a pattern we refer to as moral convergence,” Joe Hoover, a USC PhD student who led the study, told Digital Trends.

Along with Morteza Dehghani, a USC psychologist and computer scientist, the research team turned its gaze toward some 18 million tweets about the 2015 Baltimore protests against police brutality, following the death of Freddie Gray. The system scanned arrest rates, a statistic that’s often used as a proxy for violence, and found that arrests increased as “moralized” tweets increased, nearly doubling on days of violent clashes between police and protestors.

“By tracking moralized tweets posted during the 2015 Baltimore protests, we were able to observe that not only did their volume increase on days with violent protests, but also that their volume predicted hourly arrest rates, which we used as a proxy for violence, during the protests,” Hoover said. “To further unpack these effects, we conducted a series of controlled behavioral experiments and we consistently observed the same effect of moral convergence.”

From Hoover’s point of view moral convergence is a kind of fuel that contributes to violence, but it isn’t the sole factor. By identifying when a protest is more likely to turn violent, authorities may act accordingly, treating the situation with more caution.

“Protests that exhibit high convergence should be handled carefully, as even a small spark of conflict could lead to violence,” Hoover said. “Perhaps monitoring convergence on an issue could help emergency services devise and implement strategies for promoting peaceful protest.”

A paper detailing the study was published last week in the journal Nature Behavior.