GenerativeAI

andEconomic

Growth

Anewapproachtomeasuringitspotentialeconomicimpact

November2025

Note:AssessmentbasedonproductivityestimatesfocusedonGenAIthatcontinuetobechanged,especiallywithAIagentsbeingdeployed.

Contents

ExecutiveSummary 3

1.Background 5

1.1BeyondtheHype:TheImperativebehindtheGenAIImpactStudy 6

2.ChartingtheCourse:InsightsfromPastStudiesonGenAI’sImpact 7

3.TheGenAIEffect:OurAnalyticalFramework 9

3.1KPMGdynamiceconomywidemodel 9

3.2TheCGEDatabase 9

3.3OurScenarioDesignandImplementation 10

3.3.1BaselineorBusiness-As-UsualScenario 10

3.3.2RapidAdoptionofGenAI 11

3.3.3SlowAdoptionofGenAI 11

3.3.4RapidAdoptionofGenAIwithoutUp-Skilling 11

4.OurFindingsandInsights 12

4.1ImpactontheUSGrossDomesticProduct(GDP) 12

4.2WhatdoesGenAImeanfortheUSJobMarket? 14

4.2.1HowwilladoptionofGenAIimpactemploymentforvarious

laborcategories? 14

4.2.2HowwillGenAIadoptionImpactWages 15

4.3WhatIndustriesareimpactedthemostbyGenAI? 16

4.4HowwilladoptionofGenAIimpactlaborforceparticipation

andunemployment? 16

4.5WhataretheCostsofAdoptingGenAI? 2

0

4.6WhatInvestmentOpportunitiesdoesGenAIcreate? 21

4.7ImpactofGenAIontheGlobalGDP 2

3

4.8HowwilladoptionofGenAIImpactGlobalLaborMarkets? 24

5.KeyTakeaways 25

QualificationsoftheStudy 26

Appendix—A:LaborCategories 27

Appendix—B:TablesandFigures

29

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth2

updatedasgenerativeAIandagenticAIuseexpand.

ExecutiveSummary

GenerativeAI(GenAI)andagenticAIarepoisedtorevolutionizeglobaleconomiesand

employmentlandscapes.Inthepastfewyears,we’vewitnessedthelaunchoflargelanguagemodels,whichhavegrowninsophisticationandcapabilities.They’vescoredhigherona

varietyoftests,advancedfromtext-basedanswerstovoice,design,andvideo.Thisyear,thegenerativeAIhypehasshiftedtoagenticAI,whichshiftsfromansweringquestionsorcreatingcontenttocompletingtaskswithvariouslevelsofsophistication.

Withsomuchchangingalmosteverymonth,thisstudydevelopsanovelmethodfor

understandingthepotentialimpactsofGenAIadoptioninoureconomy,acrossindustries,andwithinjobroles.

Thestudyemploysadynamicglobaleconomy-widemodel,specificallytheKPMGdynamiceconomywidemodel,whichisacomputablegeneralequilibrium(CGE)model.Thismodel

capturesinter-industrylinkages,disaggregatesUShouseholdsbyincomeandrace,and

connectstheeconomyglobally.ThemodelusesthemostrecentGlobalTradeAnalysisProject(GTAP)database,widelyuseddataforeconomywidemodels.

Importantly,thestudyincorporateslaborproductivityestimatesattheindustryandsectorlevelandwithinjobrolesderivedfromtheKPMGpatent-pendingGenAIvalueassessmentmodel.

GenAIvalueassessmentisanadvancedtechnologicalservicethatoffersclientsaqualitative,outside-inassessmentofGenAIopportunitiesandthereadinessofatargetorganization.

ThisassessmentevaluatestheachievableGenAIopportunitythatanorganizationcan

potentiallycapturethroughinvestmentsinGenAI,anditalsoidentifiespeerbenchmarksforcomparativeanalysis.

Forthisstudy,weaggregatedtheGenAIopportunityvaluesbyjobroleandbytheNorth

AmericanIndustryClassificationSystem(NAICS)industry.Furthermore,basedonlocationoftheopportunityvalues,theindustriesaremappedto10globalregions.

ThismappingallowsustoaccuratelyassessthepotentialeconomicbenefitsofGenAI

implementationforspecificskillgroupswithinanindustry,suchasfinanceprofessionalsinthechemicalsindustryorhumanresourcespersonnelinthepharmaceuticalproductsindustry.

LeveragingproductivityestimatesfromlastyearthatfocusongenerativeAI—notagenticAI—thisstudyevaluatesthepotentialofAIimpactsonoureconomyusingthisnovelapproachtoinformpolicy,businessstrategy,andworkforcedevelopment.

Threescenarioswereanalyzed:abaselinescenariowithoutGenAIadoption,arapidadoptionscenario,andaslowadoptionscenario,toprojecteconomicimpactsfrom2024through2050.Inaddition,arapidadoptionscenariowithoutupskillingoflaborforcewasalsoexamined.

ToillustratehowfindingsmayevolveovertimeaslargelanguagemodelsandAIagentsalterexpectedproductivityimpacts,we’vealsoincludedananalysisofahypotheticalhighproductivityscenariowithrapidadoption.

Importantly,productivityestimateshavechangedandwillcontinuetochangewithsignificantuncertaintytodayasinnovationcontinuestodefinetheartificialintelligencespace.Aslarge

languagemodelsevolve,wewillcontinuetoevaluatehowchangesintheproductivitybenefitsderivedfrombothGenAIandagenticAIadoptionchangeoureconomy.

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth3

updatedasgenerativeAIandagenticAIuseexpand.

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth4

updatedasgenerativeAIandagenticAIuseexpand.

KeyFindings

1

2

EconomicGrowth

Ourbaselinescenariousingproductivityestimatesin2024findsthatrapidadoptionofGenAIcouldaddupto$2.84trilliontotheUSGDPby2030,and$3.37trillionby2050.Globally,theeconomycouldseeanadditionalgrowthof$11.04trillionby2050underrapidadoption,and$9.83trillionunderslowadoption.TheUSandEuropeareprojectedtogainthemost,with

$3.38trillionand$2.72trillionrespectivelyby2050.

Employment

RapidGenAIadoptionisexpectedtocreateanetgainof8.06millionjobsintheUSby2050,whilesloweradoptioncouldresultin5.79millionnewjobs.However,rapidadoptionofGenAIwithoutupskilling,couldresultinnetjoblossofaboutamillionby2050.Keysectorssuchaseducation,healthcare,andgovernmentareexpectedtoseemajorjobgrowth.

3UnemploymentRate

TherapidadoptionofGenAIisprojectedtoresultinamarginalincreaseinunemploymentrateof0.15%comparedtothebaselinein2050.Incontrast,withoutupskilling,therapidadoptionscenarioshowsanotableincreaseintheunemploymentrateto5.55%,withanetincreaseof

0.51%relativetobaseline.

4InvestmentOpportunities

BothrapidandslowadoptionofGenAIcreateconsiderableinvestmentopportunities.Rapidadoptionoffersgainsininvestmentmultipliereffectacrossmostsectorscomparedtoslowadoption.Sectorssuchaseducation,healthcare,andtransportationshowthehighestgainsfromearlyandextensiveimplementationofGenAI.

ThisstudyshowcasesthevalueofanewframeworkforevaluatingtheimpactofgenerativeandagenticAIonoureconomy.ThestudyunderscoresthetransformativepotentialofGenAIindrivingeconomicgrowth,creatingnewjobopportunities,andenhancingproductivity.

However,italsohighlightsthecriticalimportanceofupskillingandreskillingtheworkforcetofullyleveragethebenefitsofGenAIandensureequitabledistributionofitsadvantages.

StrategicinvestmentsintrainingandinfrastructurearecrucialtomaximizetheeconomicgainsfromGenAI.Policymakersandstakeholdersmustfocusoncreatinginformedpoliciesthat

addressjobdisplacement,incomeinequality,andtheneedforneweducationalandtrainingprogramstofosteraninclusiveandsustainableeconomicfuture.

Goingforward,thismodelwillhelpusbetterunderstandtheeconomicimplicationsofAIasitchangesoureconomy.

Note,findingsreflectproductivityestimatesthatcontinuetobeupdatedasgenerativeAIandagenticAIuseexpand.

GenerativeAIandEconomicGrowth5

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

GenerativeAIParadigmShift:

ShapingtheFutureofGlobalEconomiesandEmployment

1.Background

Inthespanofjustthepastcoupleofyears,GenerativeAI(GenAI)hasmovedrapidlyfrompilottrialstowidespreadadoption.ConsumersadoptedGenAI,suchasOpenAI’sChatGPTproject,attwicetherateoftabletsandsmartphones.1AndGenAIisalreadyhavinganincreasing

impactonthewayorganizationsdobusiness.It’sraretofindanenterprisetodaythatdoesn’thaveitsfootontheGenAIaccelerator.AccordingtoaQ42024KPMGAIPulseSurvey,68%ofleaderswillinvestbetween$50-$250millioninGenAIoverthenext12months,upfrom45%inQ1of2024(KPMG,2025).2TheQ22025KPMGAIPulseSurveyfoundthat90%oforganizations

werepassedexperimentingwithagenticAIandwerenowpilotinganddeploying.3ThestudyunderscoresthetransformativepotentialofAIandthepaceofinnovation.

GenAIisasubsetofartificialintelligencethatreliesongenerativemodelsthatlearnfrom

existingdataandpatternstocreatenewcontent,includingtext,images,audio,andvideo.Itcombinestechnologyanddatatodriveinnovationwithgreatspeedandefficiency—andhasfar-reachingimplicationsforhowworkgetsdoneinthefuture.GenAIisalreadyimpacting

theworkforce,with66%ofCEOssurveyedexpectingthattheimplementationofGenAIwouldrequirebothhiringnewtalent,aswellastrainingexistingtalent.GenAItechnologies,whichincludeadvancedmachinelearningmodelscapableofgeneratingtext,images,andother

media,arepoisedtorevolutionizevarioussectorsbyenhancingproductivity,creatingnewjobopportunities,anddrivinginnovation.However,understandingtheseimpactsrequiresanuancedanalysistoensurethatthebenefitsaremaximizedwhilemitigatingpotential

downsidessuchasjobdisplacementandeconomicinequality.

Globally,theeconomicimpactofGenAIextendsbeyondnationalborders,influencing

internationaltrade,labormarkets,andeconomicpolicies.Countrieswithrobusttechnologicalinfrastructuresandpoliciesthatsupportinnovationarelikelytobenefitthemost,potentiallywideningtheeconomicgapbetweendevelopedanddevelopingnations(Capraroetal.,

2024).4GenAI’seconomicimplicationsareparticularlycrucialintheUS,giventhecountry’s

leadingroleintechnologicalinnovationanditsdiverseeconomy.Theseadvancementsalso

necessitateacarefulexaminationofworkforcedynamics,asautomationmaydisplacecertainjobcategorieswhilecreatingdemandfornewskills(Hosseinietal.2024).5

DespitetherapidinvestmentinGenAI,theresearchontheeconomicimplicationsofGenAI

isstillinanascentstage.Initialfindingsindicatesubstantialpotentialeffectsonproductivity,employment,andinequality(Acemogluetal.2022).6However,thespecificsofwhere,how

andbywhomtheeconomicimpactsofGenAIwillbefeltarenotwell-understood.Bystudyingtheseimpacts,policymakersandstakeholderscanbetternavigatethecomplexitiesofGenAI,fosteringaninclusiveandsustainableeconomicfuture.

1SaraLebow,“GenerativeAIAdoptionClimbedFasterThanSmartphones,Tablets,”eMarketer(August11,2023)

2KPMG(2025a).“AIQuarterlyPulseSurvey.”Retrievedfrom:

/us/en/articles/2025/ai-quarterly-pulse-survey.html

3KPMG(2025b).“AIQuarterlyPulseSurvey.”Retrievedfrom:

AIQuarterlyPulseSurvey."Retrievedfrom:/kpmg-us/content/dam/kpmg/pdf/2025/ai-quarterly-pulse-survey-q2.pdf"

/kpmg-us/content/dam/kpmg/pdf/2025/ai-quarterly-pulse-

AIQuarterlyPulseSurvey."Retrievedfrom:/kpmg-us/content/dam/kpmg/pdf/2025/ai-quarterly-pulse-survey-q2.pdf"

survey-q2.pdf

4Capraro,V.etal.(2024).“Theimpactofgenerativeartificialintelligenceonsocioeconomicinequalitiesandpolicymaking,”PNAS

Nexus,3(6):191.Retrievedfrom:

/10.1093/pnasnexus/pgae191

5Hosseini,M.;S.P.J.M.Horbach,K.Holmes,&T.Ross-Hellauer(2024).“OpenScienceattheGenerativeAITurn:AnExploratoryAnalysisofChallengesandOpportunities.”QuantitativeScienceStudies(AdvancePublication),Retrievedfrom:

/10.1162/

qss_a_00337

6Acemoglu,D.;D.Autor,J.Hazell,&P.Restrepo(2022).“AIandJobs:EvidencefromOnlineVacancies.”NBERWorkingPaper28257,NationalBureauofEconomicResearch,(NBER)Cambridge,MA.Retrievedfrom:

/papers/w28257

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth6

updatedasgenerativeAIandagenticAIuseexpand.

1.1BeyondtheHype:TheImperativebehindtheGenAIImpactStudy

Inanerawheretechnologicaladvancementsarerapidlyreshapingourworld,understandingtheprofoundimplicationsofGenAIismorecriticalthanever.WhileGenAIcontinuesto

evolve,expandingthepotentialdisruptionsandchangestotheworkforce,thisstudy

focusesspecificallyontheimpactofproductivityasakeymetricbecauseofitsprominence.

ProductivitygainsarejustoneaspectofthebroadersetofdisruptionsandchangesthatGenAIbringstotheworkforce.TheinsightsfromthisstudyarebasedonGenAIimpacts,although

therapidemergenceofartificialgeneralintelligence(AGI)couldsoonaltertheworkforceandproductivitydynamics.7Thisresearchaimstoprovideacomprehensiveanalysisthatinformspolicy,businessstrategy,workforcedevelopment,andeconomicequity.Understandingtheseimpactsiscrucialforseveralreasons:

ShapingTomorrow—thecriticalroleofGenAIinpolicymaking:Governmentsand

policymakersneedtounderstandtheeconomicimplicationsofGenAItocreateinformedpoliciesthatbalanceinnovationwithsafetyandsecurity.Thisincludesaddressingissuessuchasjobdisplacement,incomeinequality,andtheneedforneweducationaland

trainingprograms.

StrategicEdge—howGenAIisrevolutionizingbusinessdecisions:Forbusinesses,

understandingtheeconomicimpactofGenAIcaninformstrategicdecisionsregarding

investment,workforcedevelopment,andcompetitivepositioning.Companiesthatcan

effectivelyleverageGenAIarelikelytogainsignificantcompetitiveadvantages.Thereare

considerableopportunitycosts.Theriskoffallingbehindothersandlosingtrustinyourabilitytoinnovateisrealandsignificant.Fastfolloweradvantagecontinuestoerode.Itmaybetoohardtocatchup,whichiswhyorganizationsareinvestingwithoutthetraditionalROIcase.

7ThestudyprimarilyconsidersGenAIasunderstoodbylargelanguagemodels(LLMs)atthetime,withoutagenticcapabilities.Asagenticcapabilitiesadvance,theyhavethepotentialtoaccelerateorshiftthesedynamicsconsiderably.

Future-ProofingCareers—theessentialskillsforaGenAI-drivenworld:AsGenAI

transformsindustries,therewillbeagrowingneedfornewskillsandroles.Identifyingthesechangesearlycanhelpeducationalinstitutionsandtrainingprogramsadapttopreparethefutureworkforce.

BridgingtheGap—ensuringeconomicequityintheageofGenAI:Thedistributional

impactsofGenAIcouldexacerbateexistinginequalitiesorcreatenewones.Studying

theseeffectscanhelpensurethatthebenefitsofGenAIarebroadlysharedacrossdifferentcommunitiesanddemographicgroups.

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth7

updatedasgenerativeAIandagenticAIuseexpand.

2.ChartingtheCourse

InsightsfromPastStudiesonGenAI’sImpact

GenAI,hasemergedasatransformativetechnologywithsignificanteconomicimplications.Theliteratureextensivelyexploresitsimpactonproductivityandlabormarkets.Forinstance,BrynjolfssonandLi(2024)foundthatgenerativeAItoolscansignificantlyenhanceworker

productivity,especiallyintechnicalcustomersupportroles,withanotable14%increase.

Theseproductivitygainsaremostpronouncedamonglessexperiencedcustomersupportworkers,highlightingthepotentialofAItoleveltheplayingfieldincertainjobsectors.

However,theimpactonlabormarketsiscomplex,asgenerativeAIcancreatenewjob

opportunitiesandenhanceexistingroleswhilealsoposingrisksofjobdisplacement.The

benefitstendtofavorhigher-skilledworkersandlargerfirms,leadingtounevendistributionacrosstheworkforce.ThesocioeconomicimplicationsofGenAIareequallyprofound.While

AIhasthepotentialtodemocratizeaccesstoinformationandimproveproductivity,itcanalsoexacerbateexistinginequalities.

KPMGconductedastudywiththeWorldEconomicForum(WEF)whichprovidesa

comprehensiveframeworkforunderstandingAI’simpactoneconomiesandsocieties.8

TheframeworkemphasizesAI’stransformativepotentialindrivingeconomicgrowthand

innovation,provideditisdevelopedresponsiblyandequitably.Theblueprinthighlightsthe

importanceofbuildingsustainableAIinfrastructure,curatinghigh-qualitydatasets,and

establishingethicalframeworks.Thestudyadvocatesforacollaborativeapproachinvolvinggovernments,enterprises,academia,andcivilsocietytoensureinclusiveAIdevelopment,

therebymaximizingbenefitsforall.NationsfaceuniquechallengesinfullyleveragingAI’s

benefitsduetovariousfactors.ThesechallengesincludesecuringthesubstantialinvestmentneededforadvancingAIinnovationandaddressingsignificantgapsindigitalinfrastructure,whichexacerbatestheexistingdigitaldivide.Inthiscontext,theWEFframeworkunderscorestheneedforbuildingresilientnationalAIecosystemsandpromotingwidespreadadoptionofAI,whichrequirescoordinatedmultistakeholderaction.Thestudyalsoemphasizestheneedforinternationalcollaborationtofacilitateglobaltradeandsecurecross-borderdataflows,

promotingthedevelopmentofinclusiveAImodelsandapplications.

8WorldEconomicForum(2025)“BlueprintforIntelligentEconomies:AICompetitivenessthroughRegionalCollaboration.”In

collaborationwithKPMG.Retrievedfrom:

/docs/WEF_A_Blueprint_for_Intelligent_Economies_2025.pdf

RecentliteraturesurveyrevealsthateffectivepolicymakingiscrucialtoharnessthebenefitsofAIwhilemitigatingitsadverseeffects.Theliteratureunderscorestheneedforinterdisciplinaryapproachesandrobustpolicyframeworkstoaddressthesechallenges,ensuringthatthe

advantagesofAIaremoreequitablydistributed.

TheliteratureoneconomicimpactofGenAIspansseveraldimensions,includingproductivity,labormarkets,industry-specificimpacts,andbroadereconomicgrowth.Abriefsurveyof

literatureislistedinTable2.1.

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth8

updatedasgenerativeAIandagenticAIuseexpand.

Table2.1:AReviewofLiteratureonEconomicImpactsofGenAI

Study

KeyEconomicFinding

Sector/FocusArea

Brynjolfsson&McAfee(2017)9

Productivityincreasebyupto40%

Creativeindustries

Acemoglu&Restrepo(2022)10

Displacementofupto15%ofjobs;creationofnewroles

Labormarket

Goldfarb&Tucker(2019)11

Costreductionby30%;targetingefficiencyincreaseby25%

Advertising

Agrawal,Gans,&Goldfarb(2018)12

Additional1.2%annualGDPgrowth

Broadereconomy

Bessen,J.(2022)13

NetpositivejobcreationinAI-intensivesectors

Varioussectors

Cockburn,I.,Henderson,R.,&Stern,S.(2018)14

AcceleratedinnovationcyclesandincreasedR&Dproductivity

TechnologyandR&D

Kaplan,A.,&Haenlein,M.

Enhancedcustomerengagementand

Retailandcustomer

(2019)15

personalization

service

BrynjolfssonandLi(2024)16

A14%increaseinproductivity

Technicalcustomersupportroles

Capraroetal.(2024)

Improvementinproductivity,butcouldwidenthedigitaldivideanddeepenexistinginequalities

Educationandhealthcare

9Brynjolfsson,E.,&McAfee,A.(2017).“TheBusinessofArtificialIntelligence:WhatitCan—andCannot—DoforYourOrganization.”HarvardBusinessReview.Retrievedfrom:

/2017/07/the-business-of-artificial-intelligence

10Acemoglu,D.,&Restrepo,P.(2022).“ArtificialIntelligence,Automation,&Work.”JournalofEconomicPerspectives,36(2):119-140.

11Goldfarb,A.,&Tucker,C.(2019).“DigitalEconomics.”JournalofEconomicLiterature,57(1),3-43.

12Agrawal,A.,Gans,J.,&Goldfarb,A.(2018).*PredictionMachines:TheSimpleEconomicsofArtificialIntelligence*.HarvardBusinessReviewPress.

13Bessen,J.(2022).*AIandJobs:TheRoleofDemand*.NationalBureauofEconomicResearch.WorkingPaperNo.24235.DOI:10.3386/w24235.

WhilethereisagrowingbodyofresearchonGenAI,severalgapsremainsuchasestimatingproductivitygainsacrossallsectorsandregionsfromadoptionofGenAI,linkingsector-

specificimpactswiththatofoveralleconomy,impactacrossdevelopedanddeveloping

regions,andequityimplicationsacrosstheUShouseholds.Byaddressingthesegaps,thisstudyattemptstoprovideamorecomprehensiveunderstandingoftheeconomicimpactofGenAIandhelpstakeholdersnavigatethechallengesandopportunitiesitpresents.

14Cockburn,I.,Henderson,R.,&Stern,S.(2018).*TheImpactofArtificialIntelligenceonInnovation*.NBERWorkingPaperNo.24449.DOI:10.3386/w24449.

15Kaplan,A.,&Haenlein,M.(2019).“Siri,Siri,inMyHand:Who’stheFairestintheLand?OntheInterpretations,Illustrations,andImplicationsofArtificialIntelligence.”*BusinessHorizons*,62(1),15-25.

16Brynjolfsson,E.&D.Li(2024).“TheEconomicsofGenerativeAI”.TheReporter(1).NationalBureauofEconomicResearch,Retrievedfrom:

/sites/default/files/2024-04/2024number1_0.pdf

?2025KPMGLLP,aDelawarelimitedliabilitypartnershipandamemberfirmoftheKPMGglobalorganizationofindependentmemberfirmsaffiliatedwithKPMGInternationalLimited,aprivateEnglishcompanylimitedbyguarantee.Allrightsreserved.USCS032700-1A

Note,findingsreflectproductivityestimatesthatcontinuetobeGenerativeAIandEconomicGrowth9

updatedasgenerativeAIandagenticAIuseexpand.

3.TheGenAIEffect

OurAnalyticalFramework

3.1KPMGdynamiceconomywidemodel

TheCGEmodeldevelopedinthisstudy,KPMGdynamiceconomywidemodel,isa

comprehensiveeconomicmodelthatpredictseconomy-wideimplicationsofaneconomic

shock.TheKPMGdynamiceconomywidemodelisdevelopedbasedontheextendedversionofGlobalTradeAnalysisProject(GTAP),calledMyGTAPmodel(WalmsleyandMinor,2013)17andarecursive-dynamicmodel(GTAP-RD:Aguiaretal.,2019)18enablingittocaptureand

analyzechangesovertime.Therecursivedynamicfeaturesincludeprojectingyear-over-

yearchangescapturingtheevolutionoftheeconomyincludingcapitalaccumulation,laborproductivityandotherexogenousmacro-economicforecastssuchaspopulationgrowthandchangesinglobaleconomicconditions.Thismodelcombineseconomictheoryandempiricaldatatoaccountforalltradeflowinteractionsamongindustries,consumers,andcountries

globallywhilesimultaneouslyensuringeconomicaswellasaccountingconsistency.TheKPMGdynamiceconomywidemodelprovidesacutting-edgeframeworkforanalyzingthecomprehensiveanddynamicimpactsofeconomicpolicies,providingvaluableinsightsforpolicymakersandresearchers.

3.2TheCGEDatabase

TheKPMGdynamiceconomywidemodelisbuiltuponthemostrecentGTAPdatabase

(Aguiaretal.2023),19adistinctiveandextensivelyutilizeddatabaseforeconomywide

modeling.TheGTAPdatabaseisamulti-country,multi-sectordatasetthatmapstheinput-

outputrelationshipsbetweensectorsandthetradeconnectionsbetweencountriesand

regionsglobally.Itintegratespoliciesgoverningproduction,consumption,andtrade,

drawingdatafrominternationalsourcessuchasUNCOMTRADE,theWorldBank,the

InternationalTradeCommission,andtheInternationalMonetaryFund(IMF).Thisunique,

globallybalanceddatabasealsoincludesbehavioralelasticityparametersforproducersandconsumers,someofwhichareeconometricallyestimated,whileothersarecalibratedtoalignwitheconomictheories.

ThelatestGTAPdatabaseencompasses158globalregionsand65sectorsasof2017.Althoughrunningsimulationsonfullydisaggregateddataisideal,itisnotfeasibleduetothehigh

computationalresourcesrequired.Therefore,wecreatedcustomaggregationcomprisingof

48sectors(TableB1,inAppendixB)and10globalregionstocapturetheeconomicdynamicsofthemajorglobaleconomies.20ThefactorsofproductionintheCGEdatabasetypically

includeland,labor,capital,andnaturalresources.TheInternationalLaborOrganization(ILO)

classifieslaborbyskilllevelsbasedoneducationalattainment,andthisdata