USask linguistics researcher Zhi Li, partnered with the Saskatchewan Human Rights Commission (SHRC), has been awarded almost $25,000 by the Social Sciences and Humanities Research Council of Canada to lead a team in tracking the causes of online hate on Twitter directed at Asian people.
“We will not only track the pattern and trend of anti-Asian xenophobia but also analyze the social and linguistic factors contributing to the development of such xenophobia in light of the COVID-19 pandemic,” said Li. “Ours is the first interdisciplinary data-driven study to measure this.”
Li will work with USask sociologist Hongming Cheng, and USask adjunct computer scientist Roy Ka-Wei Lee, now faculty at Singapore University of Technology and Design, as well as two graduate students and one undergraduate student.
The researchers will gather an estimated 80 million tweets, sent in Canada during the COVID-19 pandemic, from October 2019 until present.
Working with Darrell Seib, SHRC director of systemic initiatives, the research team will identify racist content in an existing database of tweets and use that information to train an algorithm to automatically pinpoint offending posts in the COVID-19 data.
“We’re looking at tweets, but also re-tweets,” said Li. “We intend to look at what makes the content ‘unfriendly’ or derogatory—the linguistic patterns behind the messages—as well as the patterns for how this information spreads.”
The researchers will study the factors that influence the spread of online hatred directed at a specific group—in this case, the Asian community.
“The importance of this type of research cannot be understated. An evidence-based understanding of online hate and discrimination will illuminate social and structural divides in society, and help human rights agencies effect change,” said David Arnot, chief commissioner of the SHRC. “Online hate cannot continue to go unchecked. The more we understand it, the better equipped we will be to respond to this kind of racial discrimination.”
The initial research will focus on the text content of hateful online tweets. In future projects, the researchers intend to comb through the same data, tracking hateful content in the form of memes, GIFs, and emojis.