AI Governance Research

We conduct AI governance research focusing on three key areas:

Our compute governance work explores how governments might control the flow of compute to effectively govern AI. In AI alignment, we investigate the intersection of Law & AI, aiming to align AI with human values by coding law into AI systems. This raises critical questions: How do we evaluate the ability of AI systems to follow the law? How do we ensure AI systems do not misinterpret the law in face of unforeseen if-then contingencies? Our work also addresses fairness in AI systems, aiming to ensure equitable outcomes across diverse populations.