1、Computing Power and theGovernance of Artifi cial IntelligenceGirish Sastry,1Lennart Heim,2Haydn Belfi eld,3Markus Anderljung,2Miles Brundage,1Julian Hazell,2,4Cullen OKeefe,1,5Gillian K.Hadfi eld,6,7Richard Ngo,1Konstantin Pilz,8George Gor,9Emma Bluemke,2Sarah Shoker,1Janet Egan,10Robert F.Trager,11

2、Shahar Avin,12Adrian Weller,13Yoshua Bengio,14Diane Coyle151OpenAI,2Centre for the Governance of AI(GovAI),3Leverhulme Centre for the Future of Intelligence,Uni.of Cambridge,4Oxford Internet Institute,5Institute for Law&AI,6University of Toronto7Vector Institute for AI,8Georgetown University,9ILINA

3、Program,10Harvard Kennedy School,11AI Governance Institute,Uni.of Oxford,12Centre for the Study of Existential Risk,Uni.of Cambridge,13Uni.of Cambridge,14Uni.of Montreal/Mila,15Bennett Institute,Uni.of CambridgeFebruary 14,2024AbstractComputing power,or compute,is crucial for the development and dep

4、loyment of artifi cial intelligence(AI)capabilities.As a result,governments and companies have started to leverage compute as a means to govern AI.Forexample,governments are investing in domestic compute capacity,controlling the flow of compute to competingcountries,and subsidizing compute access to

5、 certain sectors.However,these efforts only scratch the surface ofhow compute can be used to govern AI development and deployment.Relative to other key inputs to AI(data andalgorithms),AI-relevant compute is a particularly effective point of intervention:it is detectable,excludable,andquantifiable,a

6、nd is produced via an extremely concentrated supply chain.These characteristics,alongside the singularimportance of compute for cutting-edge AI models,suggest that governing compute can contribute to achievingcommon policy objectives,such as ensuring the safety and benefi cial use of AI.More precise