"COVID-19 was a FIFA conspiracy #curropt": An Investigation into the Viral Spread of COVID-19 Misinformation

Alexander Wang,Jerry Sun,Kaitlyn Chen,Kevin Zhou,Edward Li Gu,Chenxin Fang

DOI: https://doi.org/10.48550/arXiv.2207.01483

2022-06-13

Abstract:The outbreak of the infectious and fatal disease COVID-19 has revealed that pandemics assail public health in two waves: first, from the contagion itself and second, from plagues of suspicion and stigma. Now, we have in our hands and on our phones an outbreak of moral controversy. Modern dependency on social medias has not only facilitated access to the locations of vaccine clinics and testing sites but also-and more frequently-to the convoluted explanations of how "COVID-19 was a FIFA conspiracy"[1]. The MIT Media Lab finds that false news "diffuses significantly farther, faster, deeper, and more broadly than truth, in all categories of information, and by an order of magnitude"[2]. The question is, how does the spread of misinformation interact with a physical epidemic disease? In this paper, we estimate the extent to which misinformation has influenced the course of the COVID-19 pandemic using natural language processing models and provide a strategy to combat social media posts that are likely to cause widespread harm.

Computers and Society,Computation and Language,Social and Information Networks

What problem does this paper attempt to address?

The problem this paper attempts to address is: how to evaluate and respond to the spread of misinformation about COVID-19 on social media. Specifically, the researchers estimate the impact of misinformation on the COVID-19 pandemic through natural language processing models and propose a strategy to combat potentially harmful social media posts.

### Main Content of the Paper:
1. **Background**:
- The COVID-19 pandemic has not only brought health threats from the virus itself but also triggered a large amount of suspicious and stigmatizing information.
- Modern society's reliance on social media allows misinformation to spread rapidly, faster and wider than true information.
- This study aims to explore the interaction between the spread of misinformation and the physical pandemic and propose countermeasures.

2. **Methods**:
- **ClaimBuster**: Used to detect whether statements on social media are verifiable.
- **Tweet Legitimacy Classifier**: Used to classify the authenticity and relevance of social media content.
- **Virality Analysis**: Used to predict the potential spread of social media posts.

3. **Experiments**:
- Trained and validated using the CMU-MisCov19 dataset.
- Optimized model performance through preprocessing techniques (such as removing URLs, non-ASCII characters, etc.).
- Improved classifier accuracy using ensemble learning methods.
- Developed a binary classification model to predict whether a post will become "viral" content.

4. **Results**:
- Through a complete machine learning pipeline, it is possible to identify and classify misinformation on social media with high accuracy.
- Found that the proportion of misinformation in unpopular posts is much higher than in viral posts, indicating that users avoid interacting with misinformation to some extent.
- The proposed pipeline can serve as a practical linguistics-based misinformation detection system, helping social media platforms to timely identify and handle harmful content.

5. **Future Work**:
- Improve the ClaimBuster model to handle tweets containing multiple related statements.
- Adapt to longer social media posts to further improve model accuracy.
- Update the measurement of virality to consider factors such as the number of followers of the retweeter.
- Use hardware accelerators (such as GPUs and TPUs) to reduce runtime, allowing for more complex models to be run.

### Conclusion:
This study demonstrates the widespread dissemination of COVID-19 misinformation on social media, but the proportion of misinformation in viral posts is relatively low. The proposed pipeline can serve as a practical misinformation early warning system, helping to better filter and respond to misinformation.