Hunting Conspiracy Theories During the COVID-19 Pandemic

The fear of the unknown combined with the isolation generated by COVID-19 has created a fertile environment for strong disinformation, otherwise known as conspiracy theories, to flourish. Because conspiracy theories often contain a kernel of truth and feature a strong adversarial "other," they serve as the perfect vehicle for maligned actors to use in influence campaigns. To explore the importance of conspiracies in the spread of dis-/mis-information, we propose the usage of state-of-the-art, tuned language models to classify tweets as conspiratorial or not. This model is based on the Bidirectional Encoder Representations from Transformers (BERT) model developed by Google researchers. The classification method expedites analysis by automating a process that is currently done manually (identifying tweets that promote conspiracy theories). We identified COVID-19 origin conspiracy theory tweets using this method and then used social cybersecurity methods to analyze communities, spreaders, and characteristics of the different origin-related conspiracy theory narratives. We found that tweets about conspiracy theories were supported by news sites with low fact-checking scores and amplified by bots who were more likely to link to prominent Twitter users than in non-conspiracy tweets. We also found different patterns in conspiracy vs. non-conspiracy conversations in terms of hashtag usage, identity, and country of origin. This analysis shows how we can better understand who spreads conspiracy theories and how they are spreading them.