The NCRI documents an outbreak of Sinophobic hate emerging with COVID-19 on 4chan, an influential and extremist Web community. The NCRI deploys topic networks to expose an entire catalog of viral conspiracies and codewords around the bioweapon motif eliciting targeted ethnic hate toward Asians. The findings and methods in this case study are meant to showcase to platforms, law enforcement, intelligence, and civil-society organizations how the most innovative general methods to intersect hate on the social-cyber domain can be deployed for specific needs. The need to detect, predict and intersect emerging threats from outbreaks of weaponized information expands in tandem with COVID-19.
Using Contextus, we monitor hostility against specific identities by measuring how identity terms become targeted by hateful associations on Web communities. On over millions of comments, we witness the term ‘chink’ becoming contextually closer in meaning to ‘virus’ as the COVID-19 outbreak spreads. By contrast, terms related to Jews and derogatory terms for African Americans actually seem to show flat or decreasing relatedness to topics around the virus.
NCRI uses the seed term “chink” to examine an entire spectrum of anti-Asian propaganda and slurs. An entire cluster of Corona-related material appears to have evolved recently in the community. We discover conspiracies connecting the racial slur, ‘chink,’ to the virus. The theme of the virus being militarized as a bioweapon comprises the most widespread of the Sinophobic conspiracy theories that we encounter.
The NCRI deployed a theme-subtraction method on Contextus to create a more principled filter on Sinophobic hate as it becomes weaponized as information during the COVID-19 epidemic. When we subtract the word “chink,” from the word “bioweapon” the association of coronavirus to conspiracy is contextually nested in Sinophobichate on 4chan. We hypothesize that subtracting the hateful slur will remove all of the Sinophobic contexts from the topic network. All of the conspiracies around the coronavirus—its weaponization, use for invasion and multitude of Asian ethnic slurs—evaporate completely.
NCRI used Twitter’s verified account feed and Reddit to analyze the prevalence of the term “bioweapon” among users. Our initial results show a sharp uptick in the term corresponding with the emergence of COVID-19. While additional investigation is needed, weaponized information surfacing on mainstream platforms bears considerable risk for indoctrinating or radicalizing users.
Auto159
https://ncri.io/wp-content/uploads/NCRI-White-Paper-COVID-19-13-Apr-2020.pdf