Aidan’s research focuses on the technical aspects of AI governance. Most recently, he co-authored a research agenda for the emerging field of hardware-enabled mechanisms (HEMs) for AI governance. This field aims to design AI hardware that enforces rules on its own usage and remains robust against attempted tampering by adversaries.
Previously, Aidan conducted research on deception capabilities in large language models, trends in the development of AI tools for biology, and loopholes in US export controls on AI hardware at Center for AI Safety, Epoch, and GovAI. Before that, he led the data science operations at a venture-backed fintech startup.