Privacy preserving detection of online misinformation

Addressing online misinformation was recently highlighted as a key issue by The Royal Society in their 2021 report.

Traditionally, measuring misinformation is carried out by analysing platforms where it is found, for example crawling public posts on social media. This practice is increasingly challenging due to the wide adoption of end-to-end encryption and ethical guidelines for data processing. Since end-to-end encryption provides security benefits, interventions must be privacy-preserving by design. 

We are developing a browser-based privacy-preserving tool for misinformation detection, we will enable the detection and flagging of misinformation markers online, whilst preserving end-to-end encryption, which is essential for protecting the privacy and security of people’s online communications.

Project aims

  1. Build a browser extension tool to flag potential misinformation to online users
  2. Identify and map key misinformation techniques to linguistic features
  3. Engage with end users to understand trust in the tool and perceptions of wider issues surrounding misinformation
  4. Develop intervention training materials to improve digital information literacies