AI systems are increasingly deployed into high-stakes domains, raising urgent questions about how they should be governed, by whom, and through what combination of legal, technical, and institutional mechanisms. Governing AI requires reasoning across disciplines that rarely speak to one another: the technical characteristics of how models are trained and evaluated, the legal and regulatory instruments meant to oversee them, the economic and geopolitical forces shaping their development, and the normative questions about what we want from these systems and what we are willing to risk.
This course provides broad exposure to the central problems of AI governance. Students gain a working understanding of the technical foundations of modern AI systems and the infrastructure that supports them. The course examines the major legal and regulatory responses taking shape across the EU, US, and the rest of the world, as well as the economic and geopolitical stakes of AI, including labor markets, national security, and the geopolitics of compute. It engages the range of harms and risks that feature in AI governance debates, from privacy and fairness to existential and frontier concerns such as AGI, autonomous R&D, and CBRN. The course treats technical tools like AI evaluations as instruments of governance in their own right, and takes up some of the normative and philosophical questions raised by increasingly capable systems.
This course has no prerequisites and is designed as an introductory class, giving students broad exposure to the topics most relevant to AI governance today. It draws on perspectives from law, computer science, the social sciences, and policy, and is intended for students from any of these backgrounds. By the end, students should be able to critically read primary AI governance documents, connect the technical features of AI systems to the instruments meant to oversee them, and formulate and evaluate actionable governance proposals across national and global contexts.