Olivia
Macmillan-Scott

Final-year PhD Researcher · Computer Science

University College London · Centre for AI

I work at the intersection of technical AI research and its governance and societal implications. My current research examines reasoning and rationality in large language models, including the downstream consequences of deviations from expected behaviour. I combine a background in computer science, mathematics, and international relations with experience producing research for academic, government, and industry audiences.

AI governance frontier AI safety rationality in LLMs constitutional AI computational conflict modelling political violence game theory multi-agent systems
Olivia Macmillan-Scott
Recent news
June 2026 New preprint on bounded rationality in multi-agent systems.
Jan 2026 Winter Fellow at the Centre for the Governance of AI (GovAI), researching honesty in Constitutional AI, supervised by Atoosa Kasirzadeh.
Nov 2025 Two new preprints on cross-lingual information retrieval: generative query expansion and cross-lingual ranking.
Oct 2025 Review of current literature on Cross-Lingual Information Retrieval (CLIR) with multilingual LLMs.
Jun 2025 Started as AI Research Intern at the Alan Turing Institute, working on Cross-Lingual Information Retrieval.
2025 New paper published in PLOS Computational Biology: Maynard Smith Revisited — a multi-agent RL approach to the coevolution of signalling behaviour.