【中英文对照版】全球人工智能治理倡议.docx

全球人工智能治理倡议GlobalAIGovernanceInitiative全球人工智能治理倡议GlobalAIGovernanceInitiative（国家互联网信息办公室2023 年10月）(CyberspaceAdministrationofChina,Octoberof2023)人工智能是人类发展新领域。当 前，全球人工智能技术快速发 展，对经济社会发展和人类文明 进步产生深远影响，给世界带来 巨大机遇。与此同时，人工智能 技术也带来难以预知的各种风险 和复杂挑战。人工智能治理攸关 全人类命运，是世界各国面临的 共同课题。Artificialintelligence(AI)isanewareaofhumandevelopment.Currently,thefastdevelopmentofAIaroundtheglobehasexertedprofoundinfluenceonsocioeconomicdevelopmentandtheprogressofhumancivilization,andbroughthugeopportunitiestotheworld.However,AItechnologiesalsobringaboutunpredictablerisksandcomplicatedchallenges.ThegovernanceofAI,acommontaskfacedbyallcountriesintheworld,bearsonthefutureofhumanity.在世界和平与发展面临多元挑战 的背景下，各国应秉持共同、综 合、合作、可持续的安全观，坚 持发展和安全并重的原则，通过 对话与合作凝聚共识，构建开 放、公正、有效的治理机制，促 进人工智能技术造福于人类，推 动构建人类命运共同体。我们重中，各国应在人工智能治 理中加强信息交流和技术合作， 共同做好风险防范，形成具有广 泛共识的人工智能治理框架和标 准规范，不断提升人工智能技术 的安全性、可靠性、可控性、公 平性。我们欢迎各国政府、国际Asglobalpeaceanddevelopmentfacesvariouschallenges,allcountriesshouldcommittoavisionofcommon,comprehensive,cooperative,andsustainablesecurity,andputequalemphasisondevelopmentandsecurity.Countriesshouldbuildconsensusthroughdialogueandcooperation,anddevelopopen,fair,andefficientgoverningmechanisms,inabidtopromoteAItechnologiestobenefithumanityandcontributetobuildingacommunitywithasharedfutureformankind.WecallonallcountriestoenhanceinformationexchangeandtechnologicalcooperationonthegovernanceofALWeshouldworktogethertopreventrisks,anddevelopAIgovernanceframeworks,normsandstandardsbasedonbroadconsensus,soastomakeAItechnologiesmoresecure,reliable,controllable,andequitable.Wewelcomegovernments,internationalorganizations,组织、企业、科研院校、民间机 构和公民个人等各主体秉持共商 共建共享的理念，协力共同促进 人工智能治理。companies,researchinstitutes,civilorganizations,andindividualstojointlypromotethegovernanceofAIundertheprinciplesofextensiveconsultation,jointcontribution,andsharedbenefits.为此，我们倡议:Tomakethishappen,wewouldliketosuggestthefollowing:发展人工智能应坚持“以人 为本”理念，以增进人类共同福 祉为目标，以保障社会安全、尊 重人类权益为前提，确保人工智 能始终朝着有利于人类文明进步 的方向发展。积极支持以人工智 能助力可持续发展，应对气候变 化、生物多样性保护等全球性挑 战。Weshouldupholdapeople-centeredapproachindevelopingAI,withthegoalofincreasingthewellbeingofhumanityandonthepremiseofensuringsocialsecurityandrespectingtherightsandinterestsofhumanity,sothatAIalwaysdevelopsinawaythatisbeneficialtohumancivilization.WeshouldactivelysupporttheroleofAIinpromotingsustainabledevelopmentandtacklingglobalchallengessuchasclimatechangeandbiodiversityconservation.-面向他国提供人工智能产品 和服务时，应尊重他国主权，严 格遵守他国法律，接受他国法律 管辖。反对利用人工智能技术优 势操纵舆论、传播虚假信息，干 涉他国内政、社会制度及社会秩 序，危害他国主权。Weshouldrespectothercountries*nationalsovereigntyandstrictlyabidebytheirlawswhenprovidingthemwithAIproductsandservices.WeopposeusingAItechnologiesforthepurposesofmanipulatingpublicopinion,spreadingdisinformation,interveninginothercountries*internalaffairs,socialsystemsandsocialorder,aswellasjeopardizingthesovereigntyofotherstates.-发展人工智能应坚持“智能 向善”的宗旨，遵守适用的国际 法，符合和平、发展、公平、正 义、民主、自由的全人类共同价 值，共同防范和打击恐怖主义、 极端势力和跨国有组织犯罪集团 对人工智能技术的恶用滥用。各 国尤其是大国对在军事领域研发 和使用人工智能技术应该采取慎 重负责的态度。WemustadheretotheprincipleofdevelopingAIforgood,respecttherelevantinternationallaws,andalignAIdevelopmentwithhumanity,scommonvaluesofpeace,development,equity,justice,democracy,andfreedom.WeshouldworktogethertopreventandfightagainstthemisuseandmalicioususeofAItechnologiesbyterrorists,extremeforces,andtransnationalorganizedcriminalgroups.Allcountries,especiallymajorcountries,shouldadoptaprudentandresponsibleattitudetotheresearch,development,andapplicationofAItechnologiesinthemilitaryfield.-发展人工智能应坚持相互尊 重、平等互利的原则，各国无论 大小、强弱，无论社会制度如 何，都有平等发展和利用人工智 能的权利。鼓励全球共同推动人 工智能健康发展，共享人工智能 知识成果，开源人工智能技术。 反对以意识形态划线或构建排他 性集团，恶意阻挠他国人工智能 发展。反对利用技术垄断和单边 强制措施制造发展壁垒，恶意阻 断全球人工智能供应链。Weshouldupholdtheprinciplesofmutualrespect,equality,andmutualbenefitinAIdevelopment.Allcountries,regardlessoftheirsize,strength,orsocialsystem,shouldhaveequalrightstodevelopanduseAI.WecallforglobalcollaborationtofosterthesounddevelopmentofAI,shareAIknowledge,andmakeAItechnologiesavailabletothepublicunderopen-sourceterms.WeopposedrawingideologicallinesorformingexclusivegroupstoobstructothercountriesfromdevelopingALWealsoopposecreatingbarriersanddisruptingtheglobalAIsupplychainthroughtechnologicalmonopoliesandunilateralcoercivemeasures.-推动建立风险等级测试评估 体系，实施敏捷治理，分类分级 管理，快速有效响应。研发主体 不断提高人工智能可解释性和可 预测性，提升数据真实性和准确 性，确保人工智能始终处于人类 控制之下，打造可审核、可监 督、可追溯、可信赖的人工智能 技术。WeshouldpromotetheestablishmentofatestingandassessmentsystembasedonAIrisklevels,implementagilegovernance,andcarryouttieredandcategory-basedmanagementforrapidandeffectiveresponse.R&DentitiesshouldimprovetheexplainabilityandpredictabilityofAI,increasedataauthenticityandaccuracy,ensurethatAIalwaysremainsunderhumancontrol,andbuildtrustworthyAItechnologiesthatcanbereviewed,monitored,andtraced.逐步建立健全法律和规章制 度，保障人工智能研发和应用中 的个人隐私与数据安全，反对窃 取、篡改、泄露和其他非法收集 利用个人信息的行为。Weshouldgraduallyestablishandimproverelevantlaws,regulationsandrules,andensurepersonalprivacyanddatasecurityintheR&DandapplicationofALWeopposetheft,tampering,leaking,andotherillegalcollectionanduseofpersonalinformation.-坚持公平性和非歧视性原 则，避免在数据获取、算法设 计、技术开发、产品研发与应用Weshouldadheretotheprinciplesoffairnessandnon-discrimination,andavoidbiasesanddiscriminationbasedonethnicities,beliefs,过程中，产生针对不同或特定民族、信仰、国别、性别等偏见和nationalities,genders,etc.,duringtheprocessofdatacollection,algorithmdesign,technologydevelopment,andproductdevelopmentand歧视。application.-坚持伦理先行，建立并完善 人工智能伦理准则、规范及问责 机制，形成人工智能伦理指南， 建立科技伦理审查和监管制度， 明确人工智能相关主体的责任和 权力边界，充分尊重并保障各群 体合法权益，及时回应国内和国 际相关伦理关切。Weshouldputethicsfirst.Weshouldestablishandimproveethicalprinciples,norms,andaccountabilitymechanismsforAI,formulateAIethicalguidelines,andbuildsci-techethicalreviewandregulatorysystem.WeshouldclarifyresponsibilitiesandpowerboundariesforentitiesrelatedtoAI,fullyrespectandsafeguardthelegitimaterightsandinterestsofvariousgroups,andaddressdomesticandinternationalethicalconcernsinatimelymanner.-坚持广泛参与、协商一致、 循序渐进的原则，密切跟踪技术 发展形势，开展风险评估和政策 沟通，分享最佳操作实践。在此 基础上，通过对话与合作，在充 分尊重各国政策和实践差异性基 础上，推动多利益攸关方积极参 与，在国际人工智能治理领域形 成广泛共识。Weshouldupholdtheprinciplesofwideparticipationandconsensus-baseddecision-making,adoptagradualapproach,paycloseattentiontotechnologicaladvancements,conductriskassessmentsandpolicycommunication,andsharebestpractices.Onthisbasis,WeshouldencourageactiveinvolvementfrommultiplestakeholderstoachievebroadconsensusinthefieldofinternationalAIgovernance,basedonexchangeandcooperationandwithfullrespectfordifferencesinpoliciesandpracticesamongcountries.积极发展用于人工智能治理 的相关技术开发与应用，支持以 人工智能技术防范人工智能风 险，提高人工智能治理的技术能 力。WeshouldactivelydevelopandapplytechnologiesforAIgovernance,encouragetheuseofAItechnologiestopreventAIrisks,andenhanceourtechnologicalcapacityforAIgovernance.增强发展中国家在人工智能 全球治理中的代表性和发言权， 确保各国人工智能发展与治理的 权利平等、机会平等、规则平 等，开展面向发展中国家的国际 合作与援助，不断弥合智能鸿沟WeshouldincreasetherepresentationandvoiceofdevelopingcountriesinglobalAIgovernance,andensureequalrights,equalopportunities,andequalrulesforallcountriesinAIdevelopmentandgovernance.Effortsshouldbemadetoconductinternationalcooperationwithandprovideassistancetodevelopingcountries,tobridgethegap和治理能力差距。积极支持在联 合国框架下讨论成立国际人工智 能治理机构，协调国际人工智能 发展、安全与治理重大问题。inAIanditsgovernancecapacity.WesupportdiscussionswithintheUnitedNationsframeworktoestablishaninternationalinstitutiontogovernALandtocoordinateeffortstoaddressmajorissuesconcerninginternationalAIdevelopment,security,andgovernance.