Proponents of AI governance often look to the past for examples of how AI might be governed multilaterally. Individuals concerned about AI arms races borrow from nuclear institutions, while those who worry about uncertainty support a scientific panel akin to the IPCC. Others draw on issue areas as diverse as particle physics, civil aviation, and financial regulation.

In this talk, Dr Julia Morse, Visiting Fellow of the Oxford Martin AI Governance Initiative, will discuss the value and limitations of such approaches. Historical analogies offer an easy template for policymakers to envision AI governance. Yet each governance body reflects a distinct set of political opportunities and constraints, many of which map uneasily onto the challenge of AI. Dr Morse will highlight such considerations and offer a framework for deciding if, when, and how to borrow from historical examples when designing AI governance.