The Social Media Observatory at Royal Holloway

The vision of this fellowship is to create the Social Media Observatory at Royal Holloway University of London, to monitor and study at scale the moderation of political content within and across social media (SM) platforms, by combining social science theories with computational tools and open-source AI methods, and by putting as much emphasis on impact as on research.
Concerns about ideological, (geo)political, cultural, and algorithmic biases in SM moderation are on the rise, particularly given recent advances in black-box generative AI. However, policymakers, regulators, and academics lack the necessary data, methods, and infrastructure to tackle these challenges. The Observatory will provide a step change in the UK's capability to independently assess when and how SM platforms moderate content (or fail to do so), as well as its determinants and effects. Additionally, the Fellowship will develop a new emerging world leader and provide training to a new cohort of computational social scientists.
The Observatory will fulfil this vision with three interconnected work packages: 1. [Data & Methods] Computational pipelines to monitor SM moderation; panel surveys, data donations, and experiments to capture individual-level effects; and open-source AI models for analysing large amounts of multimodal data. 2. [Research] Observational and experimental studies on the prevalence, determinants, and effects of political SM moderation, within and across platforms. 3. [Engagement & Impact] Stakeholder engagement to translate findings into action, including partnerships with Ofcom and the UK Electoral Commission, and confirmed participation of Meta and TikTok in events and discussions.