Identifying, Measuring and Contesting Algorithmically Curated Misinformation

ABSTRACT This research examines the role of algorithms driving the online platforms in surfacing misinformation. Specifically, my dissertation work explores how can we ethically develop scalable audit pipelines to identify, measure and contest algorithmically curated misinformation. I first design experiments to audit and measure online platforms for misinformation across user features, user actions and high impact events. Next, I propose a workflow that combines human and AI capabilities to scale misinformation annotations using a value sensitive design approach. Lastly, I propose to explore how users would like to contest problematic algorithmic outputs and how can online platforms design for algorithmic contestability in scenarios where algorithms expose users to problematic content.