Skip down to main content

Countering the COVID-19 Misinfodemic with Text Similarity and Social Data Science

With Dr Scott A. Hale, and Dr Chico Camargo
24 Jun 2020
With Dr Scott A. Hale, and Dr Chico Camargo
Filming venue:

Online webinar

Misinformation about COVID-19 has led to severe harms in multiple instances: as an example, a rumor that drinking methanol would cure the virus resulted in hundreds of deaths. While end-to-end encryption is an important privacy safeguard, this encryption prevents platforms such as WhatsApp, Signal, and others from employing centralized interventions and warnings about misinformation. Several options, however, from user interface changes to tip lines to having more intelligence on client devices offer hope.

In this presentation Dr Scott A. Hale discusses how text similarity algorithms are being used to help fact-checkers locate misinformation, cluster similar misinformation, and identify existing fact-checks in the context of tip lines on platforms with end-to-end encryption. The presentation will detail research at the Oxford Internet Institute and Meedan, a global technology not-for-profit developing open-source tools for fact-checking and translation, that is actively being used by fact-checkers to improve the information available online.