应对气候危机：人工智能大有可为（英）-波士顿咨询VIP专享VIP免费

下载本文档

阅读 402
格式 pdf
大小 5.27 MB
约24页
2023-04-18
收藏
点赞(0)
海报
举报

How AI Can Be a

Powerful Tool in the Fight

Against Climate Change

July 2022

Table of

contentS

The AI Opportunity

Mitigation

Adaptation and Resilience

Fundamentals

Five Key AI Applications

Criteria for

Successful Solutions

Overcoming

Roadblocks

Beyond AI

Join Our Call

for Solutions

References

Acknowledgments

AI for the Planet Alliance:

An Introduction

Executive Summary

AI as an Essential Tool

Need for Meaningful Support

The Challenge Ahead

Report

authors

Hamid Maher, Managing Director and Partner at BCG and BCG GAMMA

Hubertus Meinecke, Managing Director and Senior Partner, Global Leader,

Climate and Sustainability Practice at BCG

Damien Gromier, Founder and Co-Chair at AI for the Planet, CEO and

Co-Founder at Startup Inside

Mateo Garcia-Novelli, Consultant at BCG

Ruth Fortmann, Senior Data Scientist at BCG GAMMA

How AI Can Be a Powerful Tool in the Fight Against Climate Change 1

AI FOR THE PLANET ALLIANCE

AN INTRODUCTION

AI for the Planet is an alliance created by Startup Inside, with Boston Consulting Group (BCG) and BCG

GAMMA as knowledge partners, in collaboration with four other organizations: the AI for Good Foundation;

the United Nations Development Programme (UNDP); the United Nations Educational, Scientic and Cultural

Organization (UNESCO); and the UN Oce of Information and Communications Technology (OICT).

This unique, multidisciplinary, and diverse coalition has several core objectives:

yPromote innovation in applying advanced analytics and articial intelligence (AI) to climate challenges,

supported by global experts from academia, startups, and the public and private sectors.

yAct as a global platform for identifying and prioritizing the leading tools and uses for AI in addressing

the climate crisis.

yIdentify and champion the most promising solutions to address climate change mitigation, as well as

adaptation and resilience, especially in the Global South, giving the solutions visibility and recognition.

yEnsure impact at scale through concrete and measurable actions such as building access to funding

and to practitioners on the ground.

yFacilitate the development of networks among project teams, investors, and experts in the eld—

including startups, corporations, and the public sector.

This alliance is looking for teams and ideas that have the

power to disrupt the climate crisis and make a meaningful

dierence by using advanced analytics and AI. We encourage

all interested parties to participate in our rst call for

solutions, including proposals at any stage of maturity if ready

for a rst pilot (at minimum) and from any sector, whether

private, public, academic, or nonprot. We will tailor support

for each solution chosen to its specic needs, ranging from

customized commercial or technical support to investor

relationships and network development.

Throughout this report, we use the term “AI” to refer to

articial intelligence and advanced analytics, dened as the

use of sophisticated data analysis techniques such as machine

learning algorithms and data engineering.

How AI Can Be a Powerful Tool in the Fight Against Climate Change 2

/24

下载本文档

文本预览下载提示常见问题

HowAICanBeaPowerfulToolintheFightAgainstClimateChangeJuly2022TableofcontentS6TheAIOpportunity8Mitigation10AdaptationandResilience12Fundamentals13FiveKeyAIApplications15CriteriaforSuccessfulSolutions16OvercomingRoadblocks18BeyondAI19JoinOurCallforSolutions20References22Acknowledgments2AIforthePlanetAlliance:AnIntroduction3ExecutiveSummary3AIasanEssentialTool4NeedforMeaningfulSupport5TheChallengeAheadReportauthorsHamidMaher,ManagingDirectorandPartneratBCGandBCGGAMMAHubertusMeinecke,ManagingDirectorandSeniorPartner,GlobalLeader,ClimateandSustainabilityPracticeatBCGDamienGromier,FounderandCo-ChairatAIforthePlanet,CEOandCo-FounderatStartupInsideMateoGarcia-Novelli,ConsultantatBCGRuthFortmann,SeniorDataScientistatBCGGAMMAHowAICanBeaPowerfulToolintheFightAgainstClimateChange1AIFORTHEPLANETALLIANCEANINTRODUCTIONAIforthePlanetisanalliancecreatedbyStartupInside,withBostonConsultingGroup(BCG)andBCGGAMMAasknowledgepartners,incollaborationwithfourotherorganizations:theAIforGoodFoundation;theUnitedNationsDevelopmentProgramme(UNDP);theUnitedNationsEducational,ScientificandCulturalOrganization(UNESCO);andtheUNOfficeofInformationandCommunicationsTechnology(OICT).Thisunique,multidisciplinary,anddiversecoalitionhasseveralcoreobjectives:yPromoteinnovationinapplyingadvancedanalyticsandartificialintelligence(AI)toclimatechallenges,supportedbyglobalexpertsfromacademia,startups,andthepublicandprivatesectors.yActasaglobalplatformforidentifyingandprioritizingtheleadingtoolsandusesforAIinaddressingtheclimatecrisis.yIdentifyandchampionthemostpromisingsolutionstoaddressclimatechangemitigation,aswellasadaptationandresilience,especiallyintheGlobalSouth,givingthesolutionsvisibilityandrecognition.yEnsureimpactatscalethroughconcreteandmeasurableactionssuchasbuildingaccesstofundingandtopractitionersontheground.yFacilitatethedevelopmentofnetworksamongprojectteams,investors,andexpertsinthefield—includingstartups,corporations,andthepublicsector.ThisallianceislookingforteamsandideasthathavethepowertodisrupttheclimatecrisisandmakeameaningfuldifferencebyusingadvancedanalyticsandAI.Weencourageallinterestedpartiestoparticipateinourfirstcallforsolutions,includingproposalsatanystageofmaturityifreadyforafirstpilot(atminimum)andfromanysector,whetherprivate,public,academic,ornonprofit.Wewilltailorsupportforeachsolutionchosentoitsspecificneeds,rangingfromcustomizedcommercialortechnicalsupporttoinvestorrelationshipsandnetworkdevelopment.Throughoutthisreport,weusetheterm“AI”torefertoartificialintelligenceandadvancedanalytics,definedastheuseofsophisticateddataanalysistechniquessuchasmachinelearningalgorithmsanddataengineering.HowAICanBeaPowerfulToolintheFightAgainstClimateChange2EXECUTIVESUMMARYThechangingclimatewillhaveamajorimpactonenvironmental,social,political,andeconomicsystemsaroundtheglobe.Wearealreadyexperiencingmanyofitsenvironmentaleffects,frommoreintensedroughtstomoredestructivestorms.Mitigationofclimatechange—alongwithadaptationandresilience—isthereforecritical,includingeffortstoachievenet-zeroemissionsby2050andtopreparefortheconsequencesofclimatechangeandminimizetheresultingharm.AIASANESSENTIALTOOLAdvancedanalyticsandartificialintelligence(apairingthatwewillrefertointhisreportsimplyas“AI”)aretoolsuniquelypositionedtohelpmanagethesecomplexissues.ABCGsurveyconductedinMay2022foundthat87%ofglobalpublic-andprivate-sectorleaderswhoareresponsibleforclimateorAItopicsbelievethatAIisausefultoolinthefightagainstclimatechange,and43%oftheseleaderssaythattheirorganizationscanenvisionusingAItohelpachievetheirclimategoals.(Seethesidebar“SurveyMethodology.”)Threeareasofapplicationareparticularlyrelevant:yMitigation.AIcanbeemployedtohelpmitigatetheclimatecrisisthroughthemeasurementofemissionsatboththemacroandmicrolevels,throughthereductionofemissionsandgreenhousegas(GHG)effects,andthroughtheremovalofexistingemissionsfromtheatmosphere.InBCG’sexperience,forexample,AIcanbeusedtohelpreduceGHGemissionsbyanamountequalto5%to10%ofanorganization’scarbonfootprint,oratotal2.6to5.3gigatonsofcarbondioxideequivalent(CO2e)ifscaledglobally.yAdaptationandResilience.AIcanbeappliedinboostingadaptationandresiliencecapacity,inpartthroughimprovedhazardprojectionsofregionalizedlong-termeffects(suchassea-levelrise)orextremeevents(suchashurricanesordroughts).Theseeffortsalsoincludethemanagementofvulnerabilityandexposure,bysuchmeansasdevelopinginfrastructurethatminimizestheimpactofclimatehazards.yFundamentals.AIcanalsobeusedtobolsterclimateresearchandmodeling;climatefinanceandanalytics;andeducation,nudging,andbehaviorchange(suchasintheformofrecommendationsforenvironmentallyfriendlypurchases).Withineachoftheseuses,AIcanbeemployedaspartoffivekeytasks:togatherandcompletecomplexdatasetsonemissions,climateimpact,andfutureprojections;tostrengthenplanninganddecisionmaking;tooptimizeprocesses;tosupportcollaborativeecosystems;andtoencourageclimate-positivebehaviors.Intheseways,itcanhelpstakeholders—frombusinessesandgovernmentstoNGOsandinvestors—pursueamoreinformedanddata-drivencourse,whileofferingthemopportunitiestocreatemeaningfulchangeinthiscriticalmoment.HowAICanBeaPowerfulToolintheFightAgainstClimateChange3NEEDFORMEANINGFULSUPPORTUserfriendlinessiscriticalifAIsolutionsaretogainwidespreadadoption,regardlessofwhethertheyaredevelopedforcorporations,governments,orthegeneralpublic.Theyneedtobereadilyaccessible,offertangiblebenefitstotheuser,andprovideclearinformationtoguideuseraction.Insomeareas—suchasemissionsmeasurementandthemonitoringofnaturalcarbonsinks—AIsolutionsarewell-establishedandreadyforbroadapplication,butmostexistingAI-relatedclimatechangesolutionsarescattered,difficulttoaccess,andlacktheresourcestoscale.Evenamongpublic-andprivate-sectorleadersengagedinclimateorAItopics,78%saythatalackofAIexpertiseisaclearobstacletotheirorganization’seffortstouseAItocombatclimatechange,77%blamealackofAIsolutions,and67%pointtoalackofconfidenceinAI-relateddataandanalysis.AIsolutionsneedmuchmoremeaningfulsupportthantheycurrentlyreceive,includingaccesstocapitalinvestments,informeddecisionmakers,andtrainedpractitioners.Inaddition,today’sAIresearch—predominantlydirectedbyresearchinstitutesandcorporationsintheGlobalNorth—risksneglectingtheneedsoftheGlobalSouth,whereclimatechangeislikelytodisproportionatelyaffectmanycountries.Accomplishingtheseendswillrequirethefollowing:yCapacityBuilding.Trainingandreskillingcanensurethatcivilservants,privatesectorleaders,andotherstakeholdersuseandinterpretAIsolutionseffectivelyincriticalcontexts.Awillingnesstoovercomevendorlock-inisalsoessentialtoensurethatuserscandeploypromisingsolutionsatscale.yResourcesandNetworks.Financialsupportcanbridgethegapbetweenacademicresearchandat-scaledeployment,andconnectionstopolicymakersandcorporateleaderscanhelpboostawarenessandadoption.Inaddition,sharingknowledgeonbestpracticesandpromisingapplicationscanhelpensurethatsolutionsarereadyforgovernmentandcorporatedeploymentatscale.AllsuchresourcesandnetworksareespeciallyvaluableforsolutionsthataddresstheneedsoftheGlobalSouth.yConfidenceandTrust.GivenAI’scomplexityandtheassociatedrisksofbiasorunethicalbehavior,whetherintentionalorunintentional,AIusersmustearntheconfidenceofleaders.Todoso,theyshouldfocusonsuchkeyareasasthequalityandgranularityoftheunderlyingdataandtheinterpretabilityoftheanalysisconducted.Despiteitspromise,AIcannotbeusedtosolvetheclimatecrisisinisolation.Rather,itisoneofmanyessentialtoolsforaddressingthisglobalchallenge.Individuals,communities,andorganizationsengagedincombatingtheclimatecrisis—regardlessoftheirformalroleinAIorclimatetopics—shouldthereforeconsiderhowotherestablishedandemergingtechnologiescanhelpandshouldassistinremovingobstaclestoscaleforthosetechnologiesaswell.Inthisreport,weprovideanoverviewofthepotentialusesofAIincombatingclimatechange,andweexaminethekeyroadblocksthatarepreventingAIsolutionsfromachievingimpactatscale.SurveyMethodologyTheoriginalquantitativefindingspresentedinthisreportarebasedonaBCGClimateAIsurveyconductedinMay2022,inwhichBCGreachedoutto1,005globalprivate-andpublic-sectorleaderswithdecision-makingauthorityonAIorclimate-changeinitiativesattheirorganizations.Mostrespondentshadauthorityoverbothtopics.Respondentsrepresentedprivate-orpublic-sectororganizationswithatleast1,000employeesoratleast$100millioninglobalannualrevenues.Thesurveysamplecovered14countries:Argentina,Australia,Brazil,Chile,China,France,Germany,India,Japan,NewZealand,SouthAfrica,Spain,theUK,andtheUS.SurveyquestionsaskedrespondentsfortheirviewsontheuseofadvancedanalyticsandAItofightclimatechange,aswellasonorganizationalobstaclestosuchefforts.HowAICanBeaPowerfulToolintheFightAgainstClimateChange4THECHALLENGEAHEADClimatechangeisalreadyhavingamajorimpactonenvironmental,social,andeconomicsystemsasglobaltemperaturesrise,droughtsworsen,andstormsbecomemoresevere.Itiscreatingmajorshiftsinphysicalrisks—fromchronicriskssuchassea-levelrisetoacuteriskssuchaswildfires—andaffectingeveryareaoftheworld.Anditisengenderingnewandcomplexchallengesforpolicymakers,corporateexecutives,andotherleaderseverywhere.IfwearetomeettheParisAgreement’sgoaloflimitingtheincreaseinaverageglobaltemperaturetolessthan2°Cabovepre-industriallevels,wemustachieverapid,deep,andinmostcasesimmediateemissionsreductionsinallsectors,accordingtoareportfromtheIntergovernmentalPanelonClimateChange(IPCC)titledClimateChange2022:MitigationofClimateChange.Suchreductionscallforsubstantialmitigationefforts,includingtheattainmentofglobalnet-zeroemissionsby2050.Themore—andthesooner—wereduceandremoveGHGemissions,thefewerchallengeswewillfaceoverthecomingdecadesaswedealwiththerisksandimpactofclimatechange.Butminimizingtheharmresultingfromclimatechangealsoentailsincreasingoureffortsatadaptationandresilienceacrossmultipletimeframes,fromimmediatecrisisresponsetolong-termplanning.Furthermore,effectiveclimateactioninsupportofmitigation,aswellasofadaptationandresilience,requirescomplementaryeffortsinresearch,finance,andeducation.ThisreportprimarilyfocusesontheUnitedNationsSustainableDevelopmentGoal(SDG)13(ClimateAction)andSDG17(PartnershipfortheGoals),astheAIforthePlanetAllianceisexpresslydedicatedtopromotingcollaborativeactiontoaddresstheclimatecrisis.Aswithmostotherclimatediscussions,however,thethemestouchedonherealludetomanyotherSDGsaswell,includingSDG11(SustainableCitiesandCommunities),SDG12(ResponsibleConsumptionandProduction),andSDG15(LifeonLand).Whereverpossible,AIforthePlanetseekstoadvanceallSDGsthroughitsactivities.TheAIforthePlanetAllianceisgratefulfortheleadershipandmomentumbuiltintheclimate-AIspacebyClimateChangeAI(CCAI),GlobalPartnershiponAI(GPAI),theCoalitionforDigitalEnvironmentalSustainability(CODES),andothers.Thisreport—likethealliance’sworkoverall—drawsoninsightsfromtheseinitiativesandattemptstoadvancethesecontributionsfurther.Inparticular,theallianceiscommittedtothethreestrategicshiftsoutlinedintheCODESactionplanlaunchedinJune2022:enablealignment,throughnetworksandsharingofbestpractices;mitigatenegativeimpacts,bypromotingresponsibleAI;andaccelerateinnovation,byhelpingbringeffectiveclimateAIsolutionstoscale.HowAICanBeaPowerfulToolintheFightAgainstClimateChange5THEAIOPPORTUNITYAIisinauniquepositionasatooltohelpmanagethesecomplexissues.InaglobalsurveyconductedbyBCGinMay2022,87%ofpublic-andprivate-sectorclimateandAIleaderssaidtheyviewAIasahelpfultoolinthefightagainstclimatechange.Furthermore,43%saidtheycanenvisionusingAIfortheirorganization’sownclimatechangeefforts—confirmingthehighlevelofinterestamongsuchleadersinthetechnology’spotentialtogeneratepositivechange.(SeeExhibit1.)BecauseofAI’suniquecapacitytogather,complete,andinterpretlarge,complexdatasetsonemissions,climateimpact,andmore,thetechnologycanbeusedtosupportallstakeholdersintakingamoreinformedanddata-drivenapproachtocombatingcarbonemissionsandbuildingagreenersociety.Itcanalsobeemployedtohelpreweightglobalclimateeffortstowardthemostat-riskregions.(Seethesidebar“AddressingtheNeedsoftheGlobalSouth.”)Assuch,theuseofAIoffersanopportunitytomakemeaningfulchangeatthiscriticalmoment,whetherthroughmitigation,throughadaptationandresilience,orthroughthefundamentalsunderpinningallclimateefforts.(SeeExhibit2.)Exhibit2-FrameworkforUsingAIinCombatingClimateChangeSources:BCGprojectexperience;ClimateChangeAI,“TacklingClimateChangewithMachineLearning”;GlobalPartnershiponAI,“ClimateChangeandAI:RecommendationsforGovernmentAction.”Note:GHG=greenhousegas;IoT=InternetofThings.Exhibit1-Public-andPrivate-SectorLeadersWhoOverseeClimateandAITopicsSupportUsingAItoFightClimateChangeSource:BCGClimateAIsurvey,May2022.Note:Allrespondentshavedecision-makingauthorityoverclimateorAItopicsattheirorganizations.Respondentswereaskediftheirorganizationhasa“clearvisionforhowadvancedanalyticsandAIcanbeusedinclimatechangeefforts.““Publicsector”excludesacademia.ofrespondentssaythatAIisahelpfultoolinthefightagainstclimatechangeRoughly40%oforganizationscanenvisionusingAIfortheirownclimateeffortsRespondents(%)87%5348444343423530HealthcareIndustrialgoodsConsumerFinancialinstitutions&insuranceAutomotivePublicsectorTechnology,media,andtelecommunicationsEnergyOverall:43%Gather,complete,andprocessdataSatelliteandIoTdataFillinggapsintemporallyandspatiallysparsedataStrengthenplanninganddecisionmakingPolicyandclimate-riskanalyticsModelingofhigher-ordereffectsBionicmanagementOptimizeprocessesSupplychainoptimizationSimulationenvironmentsSupportcollaborativeecosystemsVerticaldatasharingEnhancedcommunicationtoolsMitigationAdaptationandResilienceTopicsImprovingenergyefficiencye.g.,encouragingbehavioralchangeReducingGHGemissionsintensitye.g.,supplyforecastingforsolarenergyReducinggreenhouseeffectse.g.,acceleratingaerosolandchemistryresearchReductionEnvironmentalremovale.g.,monitoringencroachmentonforestsandothernaturalreservesTechnologicalremovale.g.,assessingcarbon-capturestoragesitesRemovalManagingcrisese.g.,monitoringepidemicsStrengtheninginfrastructuree.g.,intelligentirrigationProtectingpopulationse.g.,predictinglarge-scalemigrationpatternsPreservingbiodiversitye.g.,identifyingandcountingspeciesVulnerabilityandExposuremanagementMacro-levelmeasuremente.g.,estimatingremotecarbonnaturalstockMicro-levelmeasuremente.g.,calculatingthecarbonfootprintofindividualproductsMeasurementProjectinglocalizedlong-termtrendse.g.,regionalizedmodelingofsea-levelriseorextremeeventssuchaswildfiresandfloodsBuildingearlywarningsystemse.g.,near-termpredictionofextremeeventssuchascyclonesHazardForecastingClimateresearchandmodelinge.g.,modelingofeconomicandsocialtransitionEducation,nudging,andbehavioralchangee.g.,recommendationsforclimate-friendlyconsumptionClimatefinancee.g.,forecastingcarbonpricesSubtopicsandexamplesUsesforAIFundamentalsEncourageclimate-positivebehaviorsClimate-weightedsuggestionsClimate-friendlyoptimizationfunctionsHowAICanBeaPowerfulToolintheFightAgainstClimateChange6AddressingtheNeedsoftheGlobalSouthMorethan3billionpeopleliveinareasthatarehighlyvulnerabletoclimaterisk,suchasinSmallIslandDevelopingStates,SouthAsia,andmuchofAfrica,accordingtotheIntergovernmentalPanelonClimateChange(IPCC).Theseregionsareindangerofbeinghardesthitbyextremeweatherevents,suchasprolongeddroughtsorhurricanes,orbysea-levelrise,accordingtotheApril2022S&PGlobalreportWeatherWarning:AssessingCountries’VulnerabilitytoEconomicLossesfromPhysicalClimateRisks.Andyettheseregionshavehistoricallycontributedrelativelyfewoftheglobalemissionsthatleadtoelevatedclimaterisk.Despitetheserealities,currentclimatechangeeffortstendtoneglecttheneedsandprioritiesofvulnerableareasoftheGlobalSouth.Inpartthisreflectsthelowerconcentrationofwealthandresourcesthere,suchthattheregionlagstheGlobalNorthinadaptationandresilience.AnotherfactoristhatresearchinstitutesandcorporationsintheGlobalNorthdominateAIresearch.EvenamongAIsolutionsthatemphasizetheUNSustainableDevelopmentGoals,morethan80%arefromtheGlobalNorth,accordingtotheInternationalResearchCentreonArtificialIntelligence.ThedisproportionatefocusontheGlobalNorthriskscreatingblindspotsinAIresearchanddevelopment,anditpreventspromisingAIsolutionsfrombeingappliedinthecontextswheretheyaremosturgentlyneeded.AsGolestan(Sally)Radwan,formerAIadvisortotheEgyptianMinistryofCommunicationsandInformationTechnology,explains,“ModelsdesignedandoptimizedforGlobalNorthcountriesdonotnecessarilyapplywelltospecificconditionsinothercountries.”Asaplanet,wecannotaffordtodevelopsolutionsthatfailtoaddresstheneedsoftheGlobalSouthanddonothingtoimprovethelivesofthebillionsofpeoplelivingthere.Tothecontrary,wemustdirectAIresearchanddevelopmenttowardsolutionsthataddresstheprioritiesofpublic-andprivate-sectorleadersandpopulationsintheGlobalSouth.Inaddition,wemustensurethattheseregionshaveaccesstorelevantsolutions,alongwiththecapacitytotrainthosewhowillusethem,ifwearetoensureadoptionofthosesolutions.“Thebestapproachistobuildcapacityindifferentregions,”Radwanadds,“helpingcountriesdevelopmodelsthatworkintheirlocalcontextandexchangingknowledgeandlearningsinformedbylocalexpertise.”Thismaymeanbridginggapsindataavailabilityandmakingsolutionsdesignedforhigh-data-availabilityenvironmentsworkincontextswithlessdata.Finally,weneedtomakeglobalresourcesavailabletoensurethatAIprojectsdevelopedintheGlobalSouthreachtheirfullpotential.Similarly,connectingprojectsdevelopedintheGlobalNorthtopractitionersandusersintheGlobalSouthisacrucialstepinensuringthatthesesolutionsmeeteveryone’srequirements.HowAICanBeaPowerfulToolintheFightAgainstClimateChange7MITIGATIONOneofthemostcriticalusesofAIisinthemeasurement,reduction,andremovalofemissionsandGHGeffects.IncreasedeffortsintheseareasarecrucialtoensuringthatwemeetthegoalsoftheParisAgreement.InBCG’sexperience,forexample,AIcanbeusedtohelpgenerate5%to10%reductionsinGHGemissions,amountingtoatotal2.6to5.3gigatonsofCO2eifappliedglobally.MeasurementUnderstandingthesizeoftheproblemiscriticaltotacklingiteffectively.Measuringemissions,bothintheoverallenvironmentandatthelevelofindividualproductsandactions,allowsustotakestockofourcurrentsituationandforecastfuturetrends.Italsoallowsustoprioritizetheabatementeffortswiththehighestpotentialtoreduceemissions,atboththemacrolevelandthemicrolevel:yMacroLevel.Measurementsofoverallenvironmentalemissionsareimportantinputstomodelsthatsimulatefutureclimatescenarios.SolutionscanuseAItosupportsuchmodelsthroughmoreprecisemeasurements,forexample,orthroughtheprocessingofremote-sensingdatafromsatellites.Theycanalsouseittomonitortheimpactofnewandexistingclimatepolicies,therebyenablingteamstorefinetheirclimatemodelsandmitigationstrategies.yMicroLevel.Emissionsmeasurementsatthelevelofindividualproductsoractions—includingemissionscreatedatanypointinasupplychain—canhelpproducersunderstandtheirproducts’carbonfootprint;tracktheirprogresstowardenvironmental,social,andgovernance(ESG)targets;oridentifytheleadingwaystoabatetheirscope1,2,and3emissions.Theycanalsohelpconsumersunderstandtheemissionsthattheirownpurchasesandactionsgenerate,enablingthemtoliveacarbon-consciouslifestyleandprioritizeeffortstoreducetheircarbonfootprint.AccordingtotheCarbonDisclosureProject’sGlobalSupplyChain2021report,only38%ofcompaniesareengagingwiththeirsuppliersonclimatechange—eventhoughsupplychainscauseonaverage11.4timesasmuchemissionsascompanies’directoperations.ApplyingAItoautomatepartsofthismeasurementprocesswillspeeditupandmakeitaccessibletomorecompaniesandconsumers.ReductionRapideffortstocutthelevelofongoingemissionsandtheconsequentGHGeffects—fromoptimizingtransportationnetworkstosupportingresearchonnewtechnologies—areessentialtochangethefateoftheclimate.Weseethreecriticalopportunitiesinthisarea:.yReducingEmissionsIntensity.Althoughsomepeopleareundoubtedlywillingtoshifttolessemissions-intensivemodesoftransportation,wecannotreasonablyexpectthemtostopheatingtheirhomesinthewinter.Instead,weneedtofigureoutwaystoreducetheemissionsintensityoftheseactivities,suchasbyswitchingelectricitygenerationawayfromfossilfuels.AIsolutionscanbeusedtosupportthisswitchbyhelpingtosolvesomeofthemajoroutstandingchallengesintheenergytransition.Forexample,theymightbeemployedtoimprovesupplyforecastsforintermittentrenewableenergysources,whichDeepMindhasdemonstratedcanincreasetheeconomicvalueofwindenergyby20%.Suchforecastscanhelpreducerelianceonbatterystorageandothersourcesofpowerheldonstandby.AIsolutionsmightalsobeusedtocreatemoreefficient,real-time,electric-gridbalancing.TheymightevenbeputtoworkinsupportofR&Ddirectedtowardproducingmoreefficientsolar-panelandbatterymaterials.Onanotherfront,someindustriesarelarge-scaledirectemitters.Industrialsgenerate18%oftoday’semissions,forexample,andmuchofthatoutputcomesdirectlyfromsteel,cement,andsimilarprocesses.TeamscoulduseAItodesignmoreefficientprocessesfortheseindustriesortosupportresearchintoalternativechemicalcompositions.yReducingEmissions-GeneratingActivities.SolutionscanemployAItoreduceemissionsbyoptimizingsupplychains,suchasthroughimproveddemandforecasting(toreducewasteresultingfromoverproduction)oroptimizedtransportationofgoods(suchasbyselectingtransportationmethodsandroutesthatminimizeemissionsandcosts).ThecoreobjectiveofAItoolsintheseendeavorsistoincreasetheefficiencyofallactivities—doingmorewithless.HowAICanBeaPowerfulToolintheFightAgainstClimateChange8yReducingtheGreenhouseEffect.ActivitiesdirectedtowardreducingtheGHGeffect,referredtoas“solargeoengineering”or“solarradiationmanagement,”aimtoshieldtheEarthfromrisingtemperaturesbymodulatingtheamountofheatthatenterstheatmospherebeforeitistrappedbythegreenhouseeffect.Ifpolicymakerschoosetopursuethistypeofactivity,oneapproachmightbetoincreasetheEarth’sabilitytoreflectlightbydeployingaerosolsintheatmosphere.AIcouldbeappliedinthisinstancetoadvancethescienceofaerosols.(Forthesakeofsimplicity,thisframeworkcategorizesgeoengineeringunder“mitigation,”butitdiffersfromothermitigationactivities,suchasreducingemissions,becauseitdoesnotfocusontheconcentrationofgreenhousegasesintheatmosphere.)RemovalRemovingemissionsfromtheatmosphereisacrucialpillaroftheefforttoachieveanet-zerofuture,whetherbymonitoringnaturalcarbonsinks,suchasforests,orbysupportingtechnologicalremovalefforts,suchasdirectaircapture(DAC).AstheIntergovernmentalPanelonClimateChange’sSixthAssessmentReportonmitigationdetails,everyscenariounderwhichwesucceedinkeepingglobalwarmingbelow1.5°Cabovepre-industriallevelsreliestosomeextentoncarbonremoval.Therearetwobasiccarbonremovaloptions:yEnvironmentalRemoval.TheEarthremovescarbonfromtheatmospherebysequesteringitinnaturalsinkssuchasforests,algae,andwetlands.AI-basedsolutionscanbeusedtosupportandmoreaccuratelyquantifythisnaturalprocessbyanalyzingsatelliteimagestodetectdeforestationandestimateecosystemcarbonsequestration.Suchestimatescanhelppublic-andprivate-sectorleadersunderstandandguidethedeploymentofnaturalsolutions,includinglandmanagementandreforestationefforts.yTechnologicalRemoval.AnotherwaytoremovecarbonfromtheatmosphereisthroughscientificadvancessuchDAC,whichfiltersandcapturesCO2fromtheairasitpassesthroughamachine.ThecapturedCO2iseitherstoredunderground—insalineaquifers,forexample—orpreparedforindustrialuse.Thistechnologycurrentlyfacesquestionsregardingitsenergyefficiency,whichmaycreatebarrierstoscale,butfurtherR&Dmaybeabletoresolvetheissue.AnalternativetoDACisbioenergywithcarboncaptureandstorage(BECCS),whichgeneratesenergyoutofbiomass,includingwoodandagriculturalwaste,andcapturestheresultingCO2forstorageundergroundorforindustrialuse.SolutionscanapplyAItoassesscaptureandstoragelocations,monitorleakage,andoptimizetheindustrialprocessesandmaterialsusedforcarboncapture.Globalprivate-andpublic-sectorleadersinoursurveyseepotentialinalloftheseareasofclimate-relatedAI.Reducingandmeasuringemissionsseemtohavethegreatestpotentialbusinessvalue;about60%ofleadersseevalueinsuchusesofAI.(SeeExhibit3.)Exhibit3-Public-andPrivate-SectorLeadersSeetheGreatestBusinessValueinReducingandMeasuringEmissionsSource:BCGClimateAIsurvey,May2022.Note:Allrespondentshavedecision-makingauthorityoverclimateorAItopicsattheirorganizations.Respondentswerepermittedtogivemorethanoneanswer.Inwhichareasofclimate-relatedadvancedanalyticsandAIdoyouseethegreatestbusinessvalueforyourorganization?(%)615744423728Mitigation:reducingemissionsAdaptation:forecastinghazardsMitigation:measuringemissionsFundamentals:facilitatingclimateresearch,climatefinance,andeducationAdaptation:managingvulnerabilitiesandexposureMitigation:removingemissionsHowAICanBeaPowerfulToolintheFightAgainstClimateChange9ADAPTATIONANDRESILIENCEAdaptingtoclimatechangeisacriticalundertakingforpolicymakersandthepublic,asitboostsresiliencetotheeffectsoflong-termclimatetrendsandextremeevents.SolutionscanapplyAItopromoteadaptationandresilience,particularlyinhazardforecastingandinvulnerabilityandrisk-exposuremanagement.HazardForecastingAIiswellsuitedtohelpinforecastingandprojectingclimate-relatedhazards,whetherintheformofimprovedlong-termprojectionsofregionalizedevents,suchassea-levelrise,orintheformofearlywarningsystemsforextremeweatherevents:yProjectingRegionalizedLong-TermTrends.Accurateprojectionsofregionalizedeventscansupporteffortsbydecisionmakerstoadapttoclimatechange.Beingabletoplanaheadandknowwhichareasaremostlikelytobeaffectedcanhelpplannersdeviselong-termstrategiestosecurelocallivelihoodsandsupportsustainedeconomicgrowth.SolutionscanuseAIinthisveintocreatelarge-scaleregional-levelclimatemodels,predictcropandfishingyields,orprovidecity-levelclimateindicatorstolocalgovernments.Alternatively,decisionmakerscanemployAItoguideinvestmentsintoprotectiveinfrastructureoreconomicrestructuring.Long-termprojectionscanincludeassessmentsoftheriskofextremeevents,suchaswildfires,cyclones,orfloods—acriticalinputininfrastructureplanning.yBuildingEarlyWarningSystems.Onashortertimeframe,forecastingindividualextremeweathereventscanlimittheirdevastatingeffectsandtherebysavelives.TheuseofAIinsuchforecastshasrevolutionizedforecastingaccuracyandpermittedreal-timedataprocessingandanalysis—suchasinreviewingsatelliteimagestospotforestwildfiresandeventopredicttheirlikelihood.Forexample,in2019,UNESCObeganpartneringwithlocalexpertsinWestAfricatobuildearlywarningsystemsthatwillhelpelevencountriesmanagefloodingriskintheNigerandVoltaRiverBasins.HowAICanBeaPowerfulToolintheFightAgainstClimateChange10VulnerabilityandRisk-ExposureManagementAnotherkeycomponentofadaptationandresilienceeffortsisthestrategicmanagementofvulnerabilityandriskexposure,suchasbyplanningthedevelopmentofinfrastructureinthreatenedareassoastominimizetheimpactofclimatehazards.Forexample,sea-levelrisemaythreatenthecoastalinfrastructureofanarchipelago.Althoughaccesstowaterwaysmighttraditionallyhavebeenaneconomictradeadvantage,investingincoastalinfrastructuremightberiskiernow,makinginlandgeographicalareasmoreattractivefordevelopment.SolutionscoulduseAItohelpinthisinstancebyprojectingthelong-termeffectsofclimatechangeandsimulatingoptimalinfrastructureinvestmentsinthefaceofsuchconstraints.Inaddition,AIcanbeemployedineachofthefollowingways:yManagingCrises.Disasterresponseisacriticalaspectofvulnerabilitymanagement,enablingpublic-andprivate-sectorleaderstomakemoreinformeddecisionsundercircumstanceswheretheymightpreviouslyhavehadtooperatewithsparseorlimitedinformation.Insuchinstances,AIcanbeusedtoefficientlycombinenatural-languageprocessingwithoperationsthatinvolvegatheringandcombininginformationfrommultiplesources.Advancedanalyticscanalsobedeployedinconjunctionwithotheremergingtechnologies,suchasdrones,asseeninaprojectimplementedbyUNESCOandotherpartnerstoreducetheimpactoffloodsinTheGambia.yStrengtheningInfrastructure.Theincreasingfrequencyandintensityofnaturaldisasterscausedbyclimatechangeisputtingexistinginfrastructuretothetest.Theresilienceofhealthsystems,transportationinfrastructure,education,andinformationmediaisessentialifwearetoavoidthedirestconsequencesofnaturaldisasters.SolutionscanuseAItostrengtheninfrastructurethroughpredictivemaintenanceofstructuressuchasroads,bridges,andelectricitydistributionsystems;throughmonitoringofwaterqualityandairpollution;andeventhroughsupportfordesignprojectsandriskidentification.yProtectingPopulations.Onaverage,20millionpeopleeveryyearmustleavetheirhomesandrelocatetootherareaswithintheircountriesasaresultofextremeweatherdisasters,accordingtotheUNHCR.ResearchbyOxfamshowsthatpoorcountriesareespeciallyatrisk:80%ofthosedisplacedbyweathereventsliveinAsia,hometomorethanathirdoftheworld’spoorestpeople.SolutionscanincludeuseofAItoboosttheresilienceofcommunitiesthroughthepredictionoflarge-scalemigrationpatterns,theidentificationandmonitoringoffoodinsecurityrisk,andthehandlingofpermitallocationschemesforsocialwelfare.yPreservingBiodiversity.Humansarepartofavastecosystem.Whenanaturaldisasterstrikes,thatecosystem’sbiodiversityisputatriskofimbalanceandcollapse.AIcanbeenlistedtocontrolintelligentirrigationsystemsthatminimizetheadverseimpactofweathereventsonagriculturalproduction,toidentifycropdiseasesearly,ortopromotethetransferofknowledgebetweenbiomes,forexample.HowAICanBeaPowerfulToolintheFightAgainstClimateChange11FUNDAMENTALSAIcanbeusedtosupportresearchandeducationeffortsonclimatechange,helpingstakeholdersunderstandtherisksandimplicationsandsharewhattheylearn.Allsucheffortsreinforceandmagnifyprogresstowardmitigationandtowardadaptationandresilience.ClimateResearchandModelingModernclimateresearchandmodelinglieattheheartofallpolicydiscussionsaboutmitigationandaboutadaptationandresilience.Theyunderpinourscientificunderstandingofthefutureconsequencesofclimatechangeatthelocal,regional,andgloballevels.Withoutthem,wewouldnotbeabletoassessquantitativelyeitherthescaleofchangeinGHGemissionsorthescaleoftheadaptationeffortsneeded.AIcanbeusedtoenhancesomeaspectsofclimate-modelingresearch,inpartbecauseclimatemodelsrelyonspatiallyandtemporallysparsedatasets—amajorlimitation.Forexample,AIcanbeemployedtofillgapsinthedata,improvingthemodels’accuracyandincreasinguserconfidenceintheirresults.AIcanalsobeusedtostrengthenconnectionsinsocioeconomicandclimatemodels,suchasbycreatingfeedbackloopswithinandbetweenintegratedassessmentmodelsandclimatemodelsformoreaccuratesimulationandbetterdesignofmitigationstrategies.Evenmorepractically,AIcanbeemployedtoemulatecomputationallyexpensivesubmodelsforhigher-fidelityclimatechangesimulations,thusacceleratingthepaceofclimateresearch.Education,Nudging,andBehaviorChangeEducatingthepublicaboutthepracticalrealitiesandthreatsofglobalwarmingcanfosterchangefromthegroundup,reshapingindividuals’behavior.AIcanbeusedinthisveintohelppowerpersonalizedtoolsforestimatingcarbonfootprintsandformakingindividualrecommendationsforonlinelearningcoursesorclimate-friendlypurchases.AIisalreadyubiquitousinthedomainofgeneratingalgorithmicrecommendationsforvideostowatchorproductstobuy;similaralgorithmscanbeappliedtounderstandingbehaviorsandnudgingcitizenstowardclimate-friendlyactivities.ClimateFinanceThebroadfieldofclimatefinanceincludesinvestinginandfinancingclimate-changeinitiatives—fundingthatisessentialiftheseinitiativesaretoachievethescalenecessarytohaveameaningfulimpact.Anotherangleofclimatefinanceinvolvescarbonmarkets,whichcreatemonetaryincentivesforclimate-friendlybehaviorbyallowingfirmstotradecarbonoffsets—instrumentsthatrepresentthereduction,avoidance,orsequestrationofCO2einmetrictonunits.ClimatefinancecanbenefitfromAI’sabilitytoimprovecarbonpriceforecasts,thedynamicpricingofelectricity,andotherkeyfinancialandeconomicdatausedininvestmentandfinancingdecisions.Further,AIcanbeemployedtoimproveestimatesofthecostofclimateinaction,enablingpolicymakers,centralbankers,andprivate-sectorleaderstobetterquantifytherisksassociatedwithmovingtooslowly.HowAICanBeaPowerfulToolintheFightAgainstClimateChange1212StrengthenPlanningandDecisionMakingAIcanbeusedtoanalyzeandcomparecomplex,multilayeredproblems—forexample,whencivilservantsneedtodesignanoptimalagriculturalpolicytodealwithmorefrequentdroughtsormoresevereflooding.Theimpactofsuchpoliciesismanifoldandinterlinked,makingthisapplicationanidealuseofAI.Fortheirpart,businessleadersoftenencounterchallengestoinformeddecisionmakingbecausemethodologies,tools,datasources,andexpertisearenotavailableinatimelymanner.Asaresult,theymaynotbeabletocalculatethefinancial,operational,andsocioeconomicimpactofaparticularactionorassessthetruecostsofinaction—essentialstepsinprioritizingpotentialresponsesandallocatingresourcesefficiently.BusinessescanemployAIintheseinstancestohelpquantifybaselines,simulateoperationalchoices,andgaugeoutcomes.Whenusedinthiscapacity,AIshouldbeappliedinconjunctionwithhumandecisionmakingtoformabionicprocessinwhichhumanandAI-poweredcapabilitiesgohand-in-hand,maximizingtheeffectivenessofboth.Forexample,anAIsystemmightbeusedtoproduceanumberofpolicyscenarios,afterwhichhumanexpertscouldselectthebestalternativebasedontheirdomainexpertise.FIVEKEYAIAPPLICATIONSAcrossallofthesecriticaluses,AIcanbedirectedtowardfiveleadingapplicationsinwhichitstechnicalcapabilitiesintersectwiththeneedsofpublic-andprivate-sectordecisionmakers.Ineachcase,AIisnotthesolutionitselfbutatooltohelpleadersandcitizensmakeinformedjudgmentsabouthowtotackleclimatechallenges.(Seethesidebar“AIforShort-LivedClimatePollutants”foranillustrationofusingAIforaspecificclimatechallenge)Gather,Complete,andProcessDataAI-powereddataapproaches,suchastheInternetofThings(physicalobjectswithembeddedtechnologythatexchangedatawithdevicesorsystemsovertheinternet),giverisetonoveldataonconsumerbehaviorsandpreferences,industrialprocesses,andsupplychains.Thisdatacanbeusedtotrainalgorithmsthatuncoverbehavioralpatternsandpreferencesatscale.Moreover,AIallowstheuseofnewtypesandcombinationsofdataforanalysis,includingimages,text,andsounds.Forexample,theUNESCOWorldWaterQualityPortalforLakeChadappliesmachinelearningalgorithmstosatelliteimagerytodetectandanalyzewaterquality,generatingnewdatathatresearcherscanusetoanalyzetheeffectsofclimatechangeonwaterresources,amongotherbenefits.AIcanalsobeusedtoimputemissingdatapointsinclimate-relateddatasets—acriticalcapability,giventhatallclimate-relateddecisionsrelyonsuchdatasets.Itismuchhardertorunanalysesthatsupportgooddecisionmakingwhenkeydatasetsareincompleteorfragmented.HowAICanBeaPowerfulToolintheFightAgainstClimateChange13OptimizeProcessesAIhaslongbeenusedtosolveoptimizationproblemsinvariousdomainsacrosstheprivateandpublicsectors.Forexample,itmaybeemployedtooptimizeasupplychainbyloweringthecarbonfootprintofindividualproductsortooptimizethecoordinationofcrisisresponseteamsinthewakeofextremeweatherevents.3SupportCollaborativeEcosystemsGlobalcooperationisessentialifwearetorespondeffectivelytothechallengesposedbyclimatechange,andAI-poweredlanguagetoolscanbeusedtosupportglobalcommunicationsandcoordinationonclimategoals.Further,verticaldatasharingbetweenentitiescanbemadefasterandmoresecurethroughAI.Forexample,sharingdataonGHGemissionsbetweenfirmscancontributetoacomprehensiveviewofthecarbonfootprintofentirevaluechains—acriticalmonitoringstepforcompaniesandindividuals.AndusingAIcanhelpautomatethisreporting,imputemissingvaluesbasedonsimilarproductionprocesses,andanalyzetheinformationonawidescale.4EncourageClimate-PositiveBehaviorsAsdiscussedearlier,AIcanbeappliedinrecommendationsystemstoautomaticallyweightsuggestionstowardclimate-friendlyoptions.Similarly,AI-poweredoptimizationscanencourageclimate-friendlyprocessesbyimposingpenaltiesforalternativesthathavealargercarbonfootprint.5AIforShort-LivedClimatePollutantsContributedbyZerinOshooftheInternationalSolarAllianceRecentresearchfromtheProceedingsoftheNationalAcademyofScienceshasdemonstratedthatdecarbonizationaloneisunlikelytoachieveinternationalclimategoals.Incontrast,rapidreductionofshort-livedclimatepollutants(SLCPs)suchasblackcarbon,methane,troposphericozone,andhydrofluorocarbonstandsoutasapromisingstrategyforreducingnear-termglobalwarming,accordingtotheIntergovernmentalPanelonClimateChange(IPCC).TheIPCCnotesthatmethaneisresponsiblefor0.51°Cofthe1.06°Coftotalobservedglobalwarmingoverthepastdecadeandisalmostinextricablylinkedtomodernlife,informsrangingfromelectricitygenerationtowaste,food,clothing,andtransportation.Inresponse,AI-driventechnologyisalreadybeingusedtohelppredict,identify,andreducetheseemissions,notestheClimateChangeAIblog.Somepowerfulexampleshaveemerged:yProjectssuchasCownter,undertakenbyPicterra,andPlainsightintheagriculturalsectorhavehelpedteachcomputerstocountsheepandcattlesothatfarmerscanquantifyemissionsandpotentiallyavoidwastingmillionsofdollarsasaresultofinaccuratelivestocktallies.Similarly,adoptionofchemPEGSsensorshashelpedoptimizefertilizeruseandreducenitrousoxideemissions,accordingtoanarticlepublishedinNatureFood.yInoilandgasoperations,methaneleaksareharmfuleconomicallyandenvironmentally.MachinelearningalgorithmsdeployedbyprojectssuchasVideoGasNetenablesystemstoscancontinuouslyforinvisible,odorlessemissions,andtoquantifyleaks,peranarticleinEnergy.ResearchershaveusedAItomodelflaringfromoilandgaswellsandeventopredictleaksbeforetheyoccur.yAninternationalteamofresearchersthatincludestheLaboratoiredesSciencesduClimatetdel’Environnement,thenonprofitorganizationCarbonMapper,andtheanalyticsfirmKayrroshasappliedmachinelearningtosatelliteimagery,usingartificialneuralnetworkstoidentifypoint-of-sourceemissionsfromspace-baseddata.yBystreamliningenergy-intensiveprocesses,AI-basedtechnologycanbeusedtoreducetheircostsandlessenanyassociatedgreenhouseemissions.Asillustratedbyeachoftheseprojects,AIandmachinelearningarelikelytoplayasignificantroleinthebattleagainstSLCPemissions.HowAICanBeaPowerfulToolintheFightAgainstClimateChange14CRITERIAFORSUCCESSFULSOLUTIONSAlthoughthereareamultitudeofcriticalAIapplicationsintheclimatechangearena,anysuccessfulAIsolutionmustbeasuser-friendly,valuable,andeffectiveasastandardonlineweatherforecast.Itshouldbedesigned,scaled,andmarketedinwaysthatmakeitreadilyavailabletopractitioners.Itshouldofferenoughvaluethatitsbenefitscanbeimmediatelyperceived,potentiallychangingthewayusersthinkorbehave.Anditshouldprovideclearinformationthattheusercanacton.(SeeExhibit4.)SomeAIuses,suchasformeasuringemissionsandmonitoringnaturalcarbonsinks,arealreadyfirmlyestablishedandreadyforlarge-scaleapplication.MeasuringEmissionsAI-basedmeasurementofcarbonemissions—bothatthemacrolevelandatthelevelofindividualorganizations—hasmadeclearprogress.Examplesincludethefollowing:yClimateTRACE(TrackingReal-TimeAtmosphericCarbonEmissions),acoalitionbackedbyAlGore,usessatelliteimageryandAItomeasureglobalGHGemissions.yBCG’sCO2AIPlatformhelpscorporationsmeasure,simulate,track,andoptimizetheiremissionsatscale.MonitoringNaturalCarbonSinksVariousorganizationsaremonitoringcarbonsinksbyusingAItoanalyzesatelliteimagery—oftenincombinationwithotherdatasets—andestimateGHGemissions,sequestration,andremoval.Examplesincludethefollowing:yPachamaemployssatelliteimageryandAItomeasureandmonitorthecarbonstoredinforestsovertime,identifyinghigh-qualitycarboncreditsforitsmarketplace.yBlueSkyAnalytics,amemberofClimateTRACE,estimatesGHGemissionsfromfires,amongothersolutionsthatitoffers.Exhibit4.EffectiveAISolutionsMustMeetThreeCriteriaSource:BCGanalysis.AccessibilityThesolutionmustrelyonproventechnologythatcanbescaledandmadereadilyavailableTangiblyBeneficialThesolutionmustresolveaclearpainpointoruserneedActionabilityThesolutionmustofferrecommendationsthatareeasytoactonIllustration:weatherforecastappsWeatherforecastappsarewidelyavailableonsmartphones.UserscanquicklyaccessforecastswhenevertheywanttoUsershaveaclearneedforinformationabouttheweather—toavoiddiscomfortorworseUsersknowhowtorespondtoaweatherforecast,suchasbychoosingweather-appropriategearandavoidingexposuretodangerousweathereventsVivienneMing,co-founderandexecutivechairatSocosLabsTobetrulyuser-friendly,anewAItoolneedstosolveatangible,immediateproblemfortheuser.Eveniftoolshavecleartheoreticalvalue,themosteffectivewillhelpstakeholderstakeactiononaproblemthattheydirectlycontrolHowAICanBeaPowerfulToolintheFightAgainstClimateChange15OVERCOMINGROADBLOCKSUnfortunately,mostexistingAI-relatedclimatesolutionsarescattered,difficulttoaccess,andunderresourcedforscaling,ashighlightedbyleadingacademicexpertsinboththeCCAIreportTacklingClimateChangewithMachineLearningandtheGPAIreportClimateChangeandAI.Inaddition,someorganizationsarenotactivelyengagedinclimateandAItopics,andeventhosethatarefacesignificantroadblocks.Forexample,whenweaskedglobalpublic-andprivate-sectorleadersaboutthebiggestchallengestheirorganizationsfaceinusingAItofighttheclimatecrisis,78%citedalackofAIexpertiseinsideoroutsidetheirorganization,77%pointedtoalackofavailableAIsolutions,and67%expressedalackofconfidenceinAIdataandanalysis.(SeeExhibit5.)Exhibit5-LeadersCiteInsufficientExpertise,Availability,andConfidenceasObstaclestoUsingAIinTheirClimateChangeEffortsRespondentslistedthefollowingasobstacles(%)Source:BCGClimateAIsurvey,May2022.Note:Allrespondentshavedecision-makingauthorityoverclimateorAItopicsattheirorganizations.Respondentswerepermittedtogivemorethanoneanswer.78InsufficientaccesstoAIexpertise,insideoroutsidetheorganizationLimitedavailabilityofAIsolutionsandproducts77LackofconfidenceinAIdataandanalysis67TomaximizetheeffectivenessofAIinaddressingclimatechange,particularlyintheGlobalSouth,promisingsolutionsneedresourcesandnetworks.Theyrequiredeploymentatscalebytrainedusers.Andtheydependonextensivecapacitybuilding.Inaddition,theymustapplyAIresponsibly,usegranularandreliableunderlyingdatasets,andemphasizeresultsthatleaderscaninterpretandunderstand,iftheyaretoearnleaders’trust.HowAICanBeaPowerfulToolintheFightAgainstClimateChange16CapacityBuildingTrainingandreskillingcanhelpstakeholdersacrossthespectrum—includingcivilservants,private-sectorleaders,andothers—useandinterpretAIsolutionseffectivelyandinareasofmostcriticalneed.ManygovernmentleadersintheGlobalSouth,inparticular,haveexpressedaninterestincapacitybuilding.Thiseffortencompassesanumberofessentialelements:yTechnologicallyMindedLeaders.LeadersmustunderstandthebenefitsofAIandbewillingtoadvocateadoptingAIsolutions.yTrainedUsers.PractitionersneedtounderstandwhentoemployAIsolutionsandhowtoindependentlyimplement,manage,andtroubleshootthem;butthenecessarytrainingoftenrequiresatechnicalbackground,suchasatertiaryeducationincomputerscienceanddatascience—aprerequisitethatpreventsmanyorganizationsfromdeployingsuchsolutionsatscale.yAgileApproachestoBusinessCases.Public-andprivate-sectorleaderswhowanttosolveclimatechallenges—andinvestorswhowanttosupportpromisingAIsolutions—expecttoseeclearbusinessvaluefromanysolutionsdeployed.Demonstratingthisvaluecanbeachallenge,butitisacriticalstepinscalinganyemergingtechnology.AsMehdiGhissassi,directorandheadofproductatDeepMind,observes,“Leaders,investors,andentrepreneursalikemustbewillingtotestandadapttheirproductvisionandbusinesscasesasnewinformationcomestolight,asanAIsolutionmayprovemorevaluablewhenusedinadifferentwaythanoriginallyconceived,orforadifferentsetofstakeholders.”yPublic-PrivatePartnerships.Public-sectorsupportforemergingtechnologiessuchasAIcangreatlyspurtheirdevelopment.Forexample,nationalandinternationalresearchlaboratoriescanprovideaccesstodata,supercomputers,andotherresources,enablingentrepreneursandclimateleaderstoacceleratetheiruseofAIindevelopinginnovativesolutionsandinsights.yPrioritizationofInteroperability.Long-termcontractsandestablishedsolutionstendtocreatealock-ineffectthatmakesitdifficultforuserstoswitchtonewtechnologies,thusreducingdemandforpromisingAIapproaches.Companiesandgovernmentsshouldthereforebewillingtobuildflexibilityintotheirnegotiationswithvendors—suchasbyemphasizingdata-sharingprotocolstoensurethatnewAIsolutionscanbeintegratedmoreeasilyintocurrenttechnologies—evenifitmeanscompromisingonprice.ResourcesandNetworksPromisingAIsolutionsneedresourcesandnetworksiftheyaretoreachscaleandtobedesignedforusersworldwide,whatevertheireconomicstanding.Accesstocapital,networksandmentorship,decisionmakers,andbestpracticesiscriticaltosuccess:yCapitalInvestment.Financialsupportisessentialtobridgethegapbetweenacademicresearchanddeploymentatscale.InvestmentandgrantswillsupportpromisingAIsolutionstoscalebeyondthestartupandsmall-pilotphasestoreachindustrialornationallevels.yStrongNetworksandMentorship.MatchmakingbetweenclimateAIpractitioners(particularlyresearchersandstartups)andpotentialsupportersorusers(includingcivilservants,investors,corporates,andinternationaldevelopmentagencies)willfacilitatedevelopmentandimplementation.Engagementbetweensuchgroupswillhelppractitionersdevelopsolutionsthatfulfilltheneedsofusers.NicolaiWadstrom,founder,CEO,andGeneralPartneratBootstrapLabs,remarks,“AsaVCfirm,we’vefoundthatbuildingtherightmentorshipandnetworkstosupportentrepreneursisoftenourmostimportantactivity—sometimesbeyondthecapitalinvestmentswemake.”yAccesstoDecisionMakers.BecauseofAI’shighlytechnicalnature,public-andprivate-sectorleadersdonotalwaysapplyiteffectively,sincetheyarenotfamiliarwiththetechnology’spotentialbenefitsandlimitations.Toclosethisgap,practitionersneedaccesstopolicymakersandcorporateleaderssotheycancreateawarenessandincreaseadoption.Further,dataandcomputingpowerarenotalwaysreadilyavailableforusebyemergingsolutions,andaccesstodecisionmakerscanhelpunlocktheseresources.HowAICanBeaPowerfulToolintheFightAgainstClimateChange17ySharingBestPractices.PractitionersneedtoshareknowledgeaboutbestpracticesandpromisingusesofAIiftheyaretopreparesolutionsforwide-scalegovernmentandcorporatedeployment.Sharedlearning,suchasthroughuse-caselibraries,canhelppromisingsolutionsachievetheirfullpotentialandavoidpitfallsastheytransitionfromtheresearch,pilot,andproof-of-conceptphasestoimplementation.Best-practicesharingiscriticaltoadoptionacrossboththeGlobalNorthandGlobalSouth.ConfidenceandTrustAIpractitionersmustearntheconfidenceofleaders,giventherisksofbiasorunethicalbehaviorwhenAIisusedinappropriatelyandwithouthumanmonitoring.Consequently,practitionersmustcommitthemselveswholeheartedlytoResponsibleAI—anapproachthatencompassesappropriategovernance,thoughtfulprocesses,andtransparencyaboutAI’sroleindecisionmaking.(Seethesidebar“ResponsibleAI.”)Inaddition,practitionersshouldfocusonproducingandusingreliable,granulardatasothattheycanexplainandjustifytheresultsofanyanalysis.DoingthiswillensurethattopicexpertsseetheAItoolasavaluableadditiontotheirdecision-makingprocess.Testingtheapplicationinreal-lifecircumstancesthroughsmallpilots,aswellasconstantlyverifyingthedataandanalyses,willhelpbuildatrackrecordofaccurate,credibleresults.BEYONDAIAlthoughAIshowsgreatpromiseasawaytoaddressclimatechange,itisjustoneofmanytoolsavailabletomeetthisglobalchallenge.Likeanytechnology,ithaslimitationsandrequireseffectivedeploymentifitistoachievethedesiredresults.Further,solvingthecrisisrequiresnotjusttechnologicalinnovation,butalsothewillofdecisionmakerstotakeactionandmakethenecessarychanges—supportedinpartbyAIandotheremergingtechnologies.ResponsibleAIToensureresponsibleuseofthetechnology,AIteamsandtheiralgorithmsshouldundergoregularcomplianceaudits.Inaddition,AIteamsshouldgatherinputfromdomainexpertsandpotentiallyaffectedgroups,andtheyshouldadheretobestpracticesregardingprivacy,fairness,humancontrol,andmore.TolearnmoreaboutResponsibleAI,userscanrefertotheguidelinesco-developedbyMicrosoftandBCG.Evenifanalgorithm’sprimarypurposeisnottoaddressclimatetopics,itshouldbetrainedondatasetsthatincludeexamplesoflow-emissionprocesses,anditshouldfavoroptionsthatproduceloweremissions.IfwetrainAImodelsonhigh-emissionprocessesandusethosemodelswithoutconsideringtheGHGimpactoftheirsuggestions,wemayeventuallyundermineclimateprogressbyreplicatinghigh-carbonprocessesinthefuture.ForasystematicoverviewofAI’spotentialimpactonclimatechange—bothpositiveandnegative—weencouragereaderstorefertoAligningartificialintelligencewithclimatechangemitigation,publishedinNatureClimateChange.MoreinformationonResponsibleAIapproachestoclimatechangeisavailablefromthejointreportfromGlobalPartnershiponAI,ClimateChangeAI,andtheCentreforAI&ClimatetitledClimateChangeandAI:RecommendationsforGovernmentAction.WealsoencouragereadersrefertoSteeringAIandAdvancedICTsforKnowledgeSocieties:ARights,Openness,Access,andMulti-stakeholderPerspective,whichassessesAIthroughUNESCO’sInternetUniversalityROAMframework,andTacklingClimateChangewithMachineLearning.HowAICanBeaPowerfulToolintheFightAgainstClimateChange18JOINOURCALLFORSOLUTIONSAsclimatechangecontinuestopresentnewandcomplexchallengestoleadersaroundtheglobe,innovativeAIsolutionscanhelpmeasure,reduce,andremoveemissionsmoreeffectively;supportcommunitiesandeconomiesastheyadapttothechangingenvironmentandbecomemoreresilient;andsupportoverallclimateactionthroughresearch,climatefinance,andeducation.Althoughtherearestillnumerousroadblockstoimplementation—frominadequatefundingandtrainingtolackofaccesstodecisionmakersandbestpractices—weareexcitedtohelppushnewsolutionspastthoseroadblocksandachievetheirpotentialatscale.Weagainencourageallinterestedpartiestoparticipateinourfirstcallforsolutions,includingthoseatanystageofmaturity,givenaworkingprototype,andfromanysector.Recognizingthatallindividuals,communities,andorganizationshavearoletoplay,regardlessoftheirformalpositioninAIorclimatetopics,wefurtherinvitereaderstoconsiderhowotheremergingtechnologiescanhelpinthefightagainstclimatechangeandremoveobstaclestoscaleforthoseaswell.WealsoencouragereaderswhoareengagedintheclimateorAIcommunitiestoshareinformationontheroadblocks,frictions,andpainpointstheyfaceinengagingwithclimateanalytics,AIsolutionsfortheclimatecrisis,andotherissues.Thisvitalsupportwillinformfuturepublicationsand,moreimportantly,helpusprioritizeoureffortstoaddressthesechallenges.Formoreinformationaboutthissurveyandadditionalfindingsfromit,seetherecentBCGarticle“AIisEssentialforSolvingtheClimateCrisis”.HowAICanBeaPowerfulToolintheFightAgainstClimateChange19REFERENCESyBCG,ReduceCarbonandCostswiththePowerofAIyBCG,ResponsibleAIBuildsTrustinGovernmentyBCG,SixStepstoBridgetheResponsibleAIGapyBCG,TheTimeforClimateActionIsNowyCDP,GlobalSupplyChainReport2021yDeepMind/Google,MachineLearningCanBoosttheValueofWindEnergyyGlobalPartnershiponAIReport,ClimateChangeandAIyIPCC,AR6WGIISummaryforPolicymakersyIPCC,AR6WGIIISummaryforPolicymakersyIPCC,ClimateChange2022:MitigationofClimateChangeyIRCAI,GlobalTop100Report2021(March2022)yLynnH.Kaacketal.,AligningartificialintelligencewithclimatechangemitigationyMicrosoftxBCG,TenGuidelinesforProductLeaderstoImplementAIResponsiblyyOxfam,ForcedfromHome:Climate-FuelledDisplacementyDavidRolnicketal.,TacklingClimateChangewithMachineLearningyS&PESGResearch,WeatherWarning:AssessingCountries’VulnerabilitytoEconomicLossesfromPhysicalClimateRisksyUNESCO,SteeringAIandAdvancedICTsforKnowledgeSocietiesyUNESCO,UsingDronestoReduceImpactsofFloodsinTheGambiayUNESCO,WaterDisasterPlatformtoEnhanceClimateResilienceinAfricayUNESCO,WorldWaterQualityPortalyUNFCC,ParisAgreementyUNHCR,ClimateChangeandDisasterDisplacementHowAICanBeaPowerfulToolintheFightAgainstClimateChange20ReferencesforSidebarIIIyV.Naiketal.(2021),Chapter6:Short-livedclimateforcers,inV.Masson-Delmotteetal.(eds.),CLIMATECHANGE2021:THEPHYSICALSCIENCEBASIS,ContributionofWorkingGroupItotheSixthAssessmentReportoftheIntergovernmentalPanelonClimateChange,Masson-DelmotteV.,etal.(eds.),6–7,6–8.yG.Dreyfus,Y.Xu,D.Shindell,D.Zaelke&V.Ramanathan(2022),MitigatingClimateDisruptioninTime:ASelf-ConsistentApproachforAvoidingBothNear-TermandLong-TermGlobalWarming,PROCEEDINGSOFTHENATIONALACADEMYOFSCIENCES119(22):e21235361191yIntergovernmentalPanelonClimateChange(2021),SummaryforPolicymakers,inV.Masson-Delmotteetal.(eds.),CLIMATECHANGE2021:THEPHYSICALSCIENCEBASIS,ContributionofWorkingGroupItotheSixthAssessmentReportoftheIntergovernmentalPanelonClimateChange,FigureSPM.2yE.Lumsden(January19,2022),TheEPA’sNewMethaneReductionProposalHintsatHelpfromAI,CLIMATECHANGEAIyPicterra(2022),MachineLearningtoDetectandCountCowsfromAerialImageryyM.Grelletal.(2021),Point-of-UseSensorsandMachineLearningEnableLow-CostDeterminationofSoilNitrogen,NATUREFOOD2:981–989,981yJ.Wangetal.(2022),VideoGasNet:DeepLearningforNaturalGasMethaneLeakClassificationUsinganInfraredCamera,ENERGY238(B):1215161yV.Damodaraetal.(2020),FlarePerformanceModelingandSetPointDeterminationUsingArtificialNeuralNetworks,INTERNATIONALJOURNALOFENERGYANDENVIRONMENTALENGINEERING11:91–109,91yD.Robinson(21December2018),AIinClimateChange:MachineLearningHelpsIdentifyMethaneLeaksinGasWells,NSBUSINESSyT.Lauvauxetal.(2022),GlobalAssessmentofOilandGasMethaneUltra-Emitters,SCIENCE375(6580):557–561,557ySee,e.g.,K.Sennaar(2019),ArtificialIntelligenceforEnergyEfficiencyandRenewableEnergy–6CurrentApplications,EMERJHowAICanBeaPowerfulToolintheFightAgainstClimateChange21ACKNOWLEDGMENTSTheauthorswouldliketothanktheothermembersoftheAIforthePlanetSteeringCommitteefortheircontributions:yLambertHogenhout,UNOfficeofInformationandCommunicationsTechnologyyMarkMinevich,AIforGoodFoundationyMarielzaOliveira,UNESCOyRobertOpp,UNDPyReinaOtsuka,UNDPyPrateekSibal,UNESCOAIforthePlanetco-chairTheauthorsacknowledgethefollowingindividualswhocontributedto,reviewed,orotherwiseprovidedinputtothisreport:yBenjaminBrat,BCGGAMMAyPierreGentine,ColumbiaUniversityandLEAPyMehdiGhissassi,DeepMind/GoogleyLaurentGuengant,HitachiyStevenMills,BCGGAMMAyVivienneMing,SocosLabsyGloriaNelund,TriLincGlobalyZerinOsho,InternationalSolarAllianceyGolestan(Sally)Radwan,EgyptMinistryofCommunications&IT(former)yArunRavindran,BCGGAMMAyAtefehRiazi,MemorialSloanKetteringCancerCenterySeetYnnTan,UNESCOyEdwinvanBommel,ABNAMROBankyNicolaiWadstrom,BootstrapLabsyAliZiat,BCGGAMMATheyalsothankAliceGriffithsforwritingassistanceandStevenGrayandNadiaDannishforassistancewithediting,design,andproduction.HowAICanBeaPowerfulToolintheFightAgainstClimateChange22

1、当您付费下载文档后，您只拥有了使用权限，并不意味着购买了版权，文档只能用于自身使用，不得用于其他商业用途（如 [转卖]进行直接盈利或[编辑后售卖]进行间接盈利）。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。
3、如文档内容存在违规，或者侵犯商业秘密、侵犯著作权等，请点击“违规举报”。

碎片内容

碳中和的最新文档