Combating Online Misinformation at Scale: Developing Public Goods for Detecting and Testing Campaigns to Reduce Misinformation
Misinformation on social media platforms is rampant, from vaccine hesitancy to climate change. Policymakers and researchers regularly point to the lack of resources allocated by social media platforms to content moderation, especially in non-English languages and in the Global South, where independent fact-checking firms are also more rare. This creates a need to develop new public goods to detect misinformation on social media and to implement online interventions to reduce it. Combating misinformation could have especially high dividends on Twitter, where a few individuals influence the policy debates, attitudes and behaviors of millions – directly and indirectly through spillovers within Twitter and to other social media platforms.
We present our work that tentatively focuses on analyzing the production and diffusion of misinformation, prioritizing health and climate change; however, the techniques developed for the analysis are sector agnostic and transferable at a relatively low cost to other sectors (e.g., conflict, social cohesion). In this public repository we outline our methodolgy for detecting misinformation online to ease replication. Find out which features we have assembled and how to set up the methodology for your personal use cases. You can also report bugs here.
- Data: Datasets collected and used in the project, including fact-checked claims.
- Code: All code developed for data analysis, machine learning models, and experiment implementation.
- Models: Pre-trained models used for detecting and analyzing misinformation.
- Findings: Documentation explaining the methodologies and our analyses' findings.
This project Leveraging prior work in five countries supported by FCDO, RSB and the partnership for the SDG during the period 2019-2023. Currently, we are grateful to be supported by the World Bank's Mind, Behavior and Development Unit (eMBeD).
This project is licensed under the MIT License - see the LICENSE file for details.
The work shared here is the product of a larger team with many valued contributors: Victor Orozco, Samuel Fraiberger, Daniel Barkoczi, Ibrahim Farouq, Eaman Jahani, Blas Kolic, Karim Lasri, Hause Lin, Niyati Malhotra, Manuel Tonneau, Nina Wang, and Philipp Zimmer. In case of any questions or remarks, please feel free to reach out to us via e-mail.