Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
125 tokens/sec
GPT-4o
47 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Racism is a Virus: Anti-Asian Hate and Counterspeech in Social Media during the COVID-19 Crisis (2005.12423v2)

Published 25 May 2020 in cs.SI, cs.CL, cs.CY, cs.IR, and physics.soc-ph

Abstract: The spread of COVID-19 has sparked racism and hate on social media targeted towards Asian communities. However, little is known about how racial hate spreads during a pandemic and the role of counterspeech in mitigating this spread. In this work, we study the evolution and spread of anti-Asian hate speech through the lens of Twitter. We create COVID-HATE, the largest dataset of anti-Asian hate and counterspeech spanning 14 months, containing over 206 million tweets, and a social network with over 127 million nodes. By creating a novel hand-labeled dataset of 3,355 tweets, we train a text classifier to identify hate and counterspeech tweets that achieves an average macro-F1 score of 0.832. Using this dataset, we conduct longitudinal analysis of tweets and users. Analysis of the social network reveals that hateful and counterspeech users interact and engage extensively with one another, instead of living in isolated polarized communities. We find that nodes were highly likely to become hateful after being exposed to hateful content. Notably, counterspeech messages may discourage users from turning hateful, potentially suggesting a solution to curb hate on web and social media platforms. Data and code is at http://claws.cc.gatech.edu/covid.

Citations (153)

Summary

  • The paper introduces COVID-HATE, a dataset with over 206 million tweets and employs BERT embeddings achieving a macro-F1 score of 0.832 for reliable classification.
  • It reveals that anti-Asian hate tweets generally outnumber counterspeech responses, with a noticeable spike in counterspeech after key events like the 2021 Atlanta shooting.
  • The study demonstrates that exposure to counterspeech can curb hate propagation, providing a practical strategy for reducing online hostility.

Anti-Asian Hate and Counterspeech During the COVID-19 Crisis

The paper "Racism is a Virus: Anti-Asian Hate and Counterspeech in Social Media during the COVID-19 Crisis" investigates the propagation of anti-Asian hate speech and the role of counterspeech on Twitter amidst the COVID-19 pandemic. This paper is notable for the creation of COVID-HATE, a substantial dataset dedicated to anti-Asian commentary, compiling over 206 million tweets and a comprehensive social network consisting of more than 127 million nodes. The researchers offer a methodical examination of hate speech and counterspeech interactions and explore how these online discussions impact user behavior over time.

Study Design and Methods

The authors employ a keyword-based data collection strategy across a broad array of terms associated with both the pandemic and culturally charged phrases. Unlike rudimentary keyword searches, the paper entails a manual annotation process to distinguish specific tweet categories: anti-Asian hate, counterspeech, and neutral. The dataset eventually feeds into a text classification model utilizing BERT embeddings, achieving a macro-F1 score of 0.832, thus allowing for reliable identification of tweets falling into these categories across the dataset.

Findings

Key insights from the analysis reveal that hate tweets typically outnumber counterspeech tweets, particularly in 2020, although counterspeech saw a significant uptick following pivotal events like the 2021 Atlanta shooting. Notably, the paper identifies that while hate speech nodes are likely to mutual interact in echo chambers, a noteworthy degree of interaction between hate and counterspeech users suggests a complex web of exchanges rather than complete segregation.

Further, the analysis underscores the influence of exposure on hate propagation. Users exposed to hate tweets are more prone to produce similar content, whereas exposure to counterspeech appears to have a social inhibition effect, potentially curbing the propensity for hateful expression. In essence, counterspeech has a measurable impact that could be strategically utilized to reduce hate speech diffusion.

Practical and Theoretical Implications

The implications of this research are manifold. Practically, the paper supports the potential for employing structured counterspeech to mitigate harmful content on platforms like Twitter. The findings suggest that incorporating counterspeech strategies could be an effective non-invasive method to address online hate, with promise for broader application in combating various forms of digital intolerance.

Theoretically, the paper contributes significantly to the understanding of social media dynamics and hate speech diffusion, providing a nuanced perspective on the role of counterspeech. Future research could expand upon this work by exploring counterspeech in other languages and cultural contexts or by integrating real-world event data to fully comprehend the interplay between offline events and online discourse.

In sum, this research offers vital insights into how the digital landscape can be better managed to foster a less hostile and more inclusive virtual environment during challenging times exacerbated by unprecedented global crises like COVID-19.

Youtube Logo Streamline Icon: https://streamlinehq.com