2024全球人工智能挑战框架公约：加快国际合作以确保AI有益安全包容发展 英文版 .docxVIP

FrameworkConventionCI

FrameworkConvention

onGlobalAIChallenges

Acceleratinginternationalcooperationtoensurebeneficial,safeandinclusiveAI

DuncanCass-BeggsStephenClareDawnDimowoZaheedKara

AbouttheGlobalAIRisksInitiative

TheGlobalAIRisksInitiativeattheCentreforInternationalGovernanceInnovationwascreatedtoadvancetheinternationalgovernancethatwillbeneededtomanageglobalAIrisks.TheInitiativeaimstomobilizetheresources,talentandinfluenceofpolicymakers,AIresearchers,governanceexpertsandcivilsocietytoreduceglobalrisksfromadvancedAIsystems.ItseekstobuildunderstandingoftheimportanceofglobalrisksfromAIandaccesstoworkablepolicyoptionstomitigatetheseriskssuccessfully.

ThisdiscussionpaperproposesthedevelopmentofaninternationalFrameworkConventiononGlobalAIChallengesaccompaniedbyspecificsupportingprotocolstoaddressspecificglobalchallengesraisedbyadvancedAI.Thepaperdrawsonawealthofexistingresearchandpolicyefforts,aswellasvaluablediscussionsandfeedbackfrommanyquarters.Nevertheless,therecommendationsarepreliminaryandintendedtosupportfurtherdialogue,reflectionandaction.Theauthorswelcomefeedback,aswellassuggestedimprovementsandcollaborations.Theymaybereachedat

globalairisks@.

Credits

ExecutiveDirector,GlobalAIRisksInitiativeDuncanCass-Beggs(leadauthor)SeniorResearchAssociate,GlobalAIRisksInitiativeStephenClare(leadauthor)ProgramManager,GlobalAIRisksInitiativeDawnDimowo

SeniorResearchAssociate,GlobalAIRisksInitiativeZaheedKaraSeniorPublicationsEditorJenniferGoyder

PublicationsEditorSusanBubak

GraphicDesignerSamiChouhdary

Copyright?2024bytheCentreforInternationalGovernanceInnovation

TheopinionsexpressedinthispublicationarethoseoftheauthorsanddonotnecessarilyreflecttheviewsoftheCentreforInternationalGovernanceInnovationoritsBo