Should Artificial Intelligence Governance be Centralised? Design Lessons from History

Can effective international governance for artificial intelligence remain\nfragmented, or is there a need for a centralised international organisation for\nAI? We draw on the history of other international regimes to identify\nadvantages and disadvantages in centralising AI governance. Some\nconsiderations, such as efficiency and political power, speak in favour of\ncentralisation. Conversely, the risk of creating a slow and brittle institution\nspeaks against it, as does the difficulty in securing participation while\ncreating stringent rules. Other considerations depend on the specific design of\na centralised institution. A well-designed body may be able to deter forum\nshopping and ensure policy coordination. However, forum shopping can be\nbeneficial and a fragmented landscape of institutions can be self-organising.\nCentralisation entails trade-offs and the details matter. We conclude with two\ncore recommendations. First, the outcome will depend on the exact design of a\ncentral institution. A well-designed centralised regime covering a set of\ncoherent issues could be beneficial. But locking-in an inadequate structure may\npose a fate worse than fragmentation. Second, for now fragmentation will likely\npersist. This should be closely monitored to see if it is self-organising or\nsimply inadequate.\n