Skip to main content
About

About

PhD topic: AI-driven polarisation: a computational approach

PhD supervisors: Dr Jens Koed Madsen, Professor Alex Gillespie, Dr Lee de Wit (Cambridge)

Sonja is a PhD candidate, studying how user psychology, recommender systems, and AI-generated content interact to drive belief polarisation in online information environments. By combining empirical and computational approaches – such as agent-based models – her work explores how emerging technologies, like LLMs, may threaten epistemic security. Her broader research interests lie at the intersection of cognitive psychology, technology, and security, particularly how to design and regulate information ecosystems, to make them resilient to AI-driven manipulation.

Prior to her PhD, Sonja completed an MPhil in Psychology at the University of Cambridge, supervised by Dr Lee de Wit and Dr David Young. Using Bayesian belief networks, her dissertation examined why we observe greater belief polarisation for social policy issues compared to economic ones. Alongside academic work, while an undergraduate in the PBS department at LSE, Sonja was President of LSE's Behavioural and Psychological Science Society, where she co-launched the UK's first inter-university Behavioural Science Careers Fair with UCL BIS and Warwick BIT, and co-founded Irrationale, LSE's student psychology journal.

Awards

  • LSE PhD Studentship
  • Best Performance in Programme, LSE BSc Psychological and Behavioural Science 2024