非洲人工智能：產(chǎn)生影響的用例（英）-78正式版-WN8

AIforAfrica:

Usecases

delivering

impact

July2024

Copyright?2024GSMA

GSMAGSMACentralInsightsUnit

TheGSMAisaglobalorganisationunifyingthemobile

ecosystemtodiscover,developanddeliverinnovation

foundationaltopositivebusinessenvironmentsand

societalchange.Ourvisionistounlockthefullpowerof

connectivitysothatpeople,industry,andsocietythrive.

Representingmobileoperatorsandorganisationsacrossthe

mobileecosystemandadjacentindustrieschallenges,the

GSMAdeliversforitsmembersacrossthreebroadpillars:

ConnectivityforGood,IndustryServicesandSolutions,and

Outreach.Thisactivityincludesadvancingpolicy,tackling

today’sbiggestsocietal,underpinningthetechnologyand

interoperabilitythatmakemobilework,andprovidingthe

world’slargestplatformtoconvenethemobileecosystem

attheMWCandM360seriesofevents.

TheCentralInsightsUnit(CIU)sitsatthecoreofGSMA

MobileforDevelopment(M4D)andproducesin-depth

researchontheroleandimpactofmobileanddigital

technologiesinadvancingsustainableandinclusive

development.TheCIUengageswithpublicandprivate

sectorpractitionerstogenerateuniqueinsightsandanalysis

onemerginginnovationsintechnologyfordevelopment.

Throughourinsights,wesupportinternationaldonorsto

buildexpertiseandcapacityastheyseektoimplement

digitisationinitiativesinlow-andmiddle-incomecountries

throughpartnershipswithinthedigitalecosystem.

Contactusbyemail:centralinsights@

Weinviteyouto?ndoutmoreat

FollowtheGSMAonTwitter/X:@GSMA

ThisinitiativehasbeenfundedbyUKAidfromtheUK

GovernmentandissupportedbytheGSMAandits

members.Theviewsexpresseddonotnecessarilyre?ect

theUKGovernment’so?cialpolicies.

AxumisanAfrocentricimpactcompanydedicatedto

fosteringclimate-positive,digitallyinnovative,inclusive

growthacrossAfrica,theMiddleEastandaroundthe

world.Thecompanyworkswithlocallyandglobally

in?uentialleadersthatseektodrivesustainable

development,inclusionandprosperity.

Throughstrategicleadershipandtheabilitytotransform

ideasintoreality,Axumpartnerswithdiversestakeholders

todrivepositivechange.Boastingover150yearsof

collectiveleadershipexperience,andateamofnearly

100across10o?cesinAfricaandtheMiddleEast,Axum

leveragesawealthofmultisectoralandmulticultural

expertisetohelpclientsnavigatepressingglobal

challengesandrealiseAfricaandtheMiddleEast’s

immensepotential.

Authorsandcontributors

Authors:EugénieHumeauandTanviDeshpande

Contributor:DanieleTricarico

Acknowledgements

ThisreportdrawsonresearchconductedfortheGSMAby

Axum.WewouldliketothankGathoniKang'ethe,Jamila

Raji,JonathanMunge,SalmaAitHssayene,IsisNyong’o

MadisonandRobinMillerfortheircontribution.

Wewouldliketothankthefollowingindividualswhowere

partofourExpertAdvisoryGroupandprovidedguidance

andexpertiseduringtheresearchprojectthroughvarious

engagements:AlbanOdhiambo(TonyBlairInstitutefor

GlobalChange),DeshniGovender(GIZ–FAIRForward),

DrGirmawAbebeTadesse(MicrosoftAIforGoodLab),

KateKallot(Amini),KoliweMajama(MozillaFoundation),

LavinaRamkissoon(AfricanUnion),LilySteele(Global

InnovationFund),LinetKwamboka(GlobalPartnershipfor

SustainableDevelopmentData),LukasBorkowski(Viamo),

MatthewSmith(IDRC)andDrOlubayoAdekanmbi(Data

ScienceNigeria).

WewouldalsoliketothankDrEmmyChirchirandDr

EmmelineSkinner(FCDOEastAfricaResearchand

InnovationHub),KristinKlose(FCDOSouthAfricaResearch

andInnovationHub)andOluwasegunAdetunde(FCDO

WestAfricaResearchandInnovationHub)fortheirinput

andfeedback.

Finally,wewanttothankthemanyindividualsand

organisationsthatcontributedtotheresearch.Afulllistof

organisationsconsultedfortheresearchislistedattheend

ofthereport.

Contents

Executivesummary4

1.Introduction7

2.Researchobjectivesandmethodology10

3.De?ningAI14

WhatisAI?15

AIfundamentalsinAfrica16

4.Usecasesdeliveringimpact29

Keytrendsacrossusecases30

Agricultureandfoodsecurity35

Energy42

Climateaction49

5.Towardsathrivingecosystem56

Creatingaconducivepolicyenvironment57

Fosteringpartnerships60

Unlocking?nancingatscale63

Supportingresearchanddevelopment64

6.Conclusionandrecommendations66

Annexes72

Listof?gures

Figure1EstimatedannualvalueoftheAImarket

inAfricarelativetotheglobalmarket

Figure2PotentialvalueaddedbyAItothe

Africaneconomy

Figure3TheAIecosystemframework

Figure14dAllocationofusecasesbyownership

Figure14eAllocationofusecasesbytypeof

solution

Figure15Agriculture’scontributiontoGDPand

labourforcebycountry

Figure4The?veVsofbigdata

Figure5Examplesofdatatypesandsourcesfor

AIfordevelopment

Figure6Generationandusageofdatasets

globally

Figure16Overviewofusecasesinagricultureand

foodsecurity

Figure17Heatmapofusecasesinagricultureand

foodsecuritybycountry

Figure18Accesstoelectricitybycountry

Figure7Prevalenceofinternetcontentin

Africanlanguagescomparedtoglobal

benchmarks

Figure19Africancountrieswithmorediesel

generatorcapacitythangridcapacity

Figure20ElectricitygridmixinSub-SaharanAfrica

Figure8AIinfrastructureandcomputelayers

Figure21Overviewofusecasesinenergy

Figure9Currentandprojectedsmartphone

adoptionbycountry

Figure22Heatmapofusecasesinenergyby

country

Figure10Projectedpercentageof5Gconnections

bycountry

Figure11SkillsetsrequiredbyAIbuildersandAI

users

Figure12Whatmakesagoodprompt?

Figure13Distributionofusecaseapplicationsby

country

Figure27AIpolicydevelopmentinAfrica

Figure14aAllocationofusecasesbysector

Figure14bAllocationofusecasesbytypeofAI

Figure14cAllocationofusecasesbytypeof

organisation

Figure23CO2emissionsbyregion

Figure24Overviewofusecasesinclimateaction

Figure25Remotesensingasatooltosupport

climateaction

Figure26Heatmapofusecasesinclimateaction

bycountry

Figure28Typesofactorsinvolvedinpartnerships

forAI

Listoftables

Table1Researchmethodology

Table2Hungerassessmentbycountry

Table5VenturecapitalinvestmentsinAIby

country

Table3Vulnerabilitytoclimatechangeand

readinesstoimproveresilience

Table4Venturecapitalinvestmentsintechby

country

Table6CountryranksforR&Dcapabilities

Table7KeyrecommendationstosupportAI

deploymentandadoption

Listofboxes

Box1Buildinglocallanguagedatasets:Challenges

andopportunities

Box5Usecasedeepdive:Foodsecurity

forecasting

Box2Whatarethebene?tsofedgecomputing?

Box3'AskViamoAnything'bringsgenerative

AItechnologytodigitallydisconnected

communities

Box4Usecasedeepdive:Precisionagriculture

Box6Usecasedeepdive:Energyaccessand

demandassessment

Box7Usecasedeepdive:Biodiversitymonitoring

Box8UNESCO’shumanrights-centredapproachto

theethicsofAI

AIforAfrica:Usecasesdeliveringimpact

2/76

Listofacronyms

AIArti?cialIntelligenceIVRInteractiveVoiceResponse

CDRCallDetailRecordsLLMLargeLanguageModel

DFSDigitalFinancialServicesMLMachineLearning

EWSEarlyWarningSystemMNOMobileNetworkOperator

GDPGrossDomesticProductNLPNaturalLanguageProcessing

GISGeographicInformationSystemNRMNaturalResourcesManagement

GPTGenerativePre-trainedTransformerPAYGPay-As-You-Go

GPUGraphicProcessingUnitPPPPublic-PrivatePartnership

HPCHighPerformanceComputingR&DResearchandDevelopment

HWCHuman-WildlifeCon?ictSHSSolarHomeSystem

IoTInternetofThingsUSSDUnstructuredSupplementaryServiceData

De?nitions

AIforDevelopment:Weusetheterm‘AIfor

development’torefertotheuseofAIand

itsapplicationswiththepotentialtoaddress

developmentchallengesinlow-andmiddle-income

countries.

Algorithm:Aprocessorsetofrulestobefollowed

incalculations,especiallybyacomputer,tosolvea

problem.

Arti?cialintelligence:Arti?cialintelligence(AI)

iscomprisedofwidelydi?erenttechnologiesthat

canbebroadlyde?nedas“self-learning,adaptive

systems.”1AIhasthecapabilitytounderstand

language,solveproblems,recognisepicturesand

learnbyanalysingpatternsinlargesetsofdata.

BigTech:Inthisreport,BigTechplayersreferto

thelargetechcompaniesknownglobally,including

Google,Microsoft,IBM,Meta,andAmazon.The

terms'BigTech'and'largetechcompanies'areused

interchangeablyinsomecontexts.

Computervision:AtypeofAIthatenables

computersandothermachinestoidentifyand

interpretvisualinputsfromimagesandvideos.3

GenerativeAI:AtypeofAIthatinvolvesgenerating

newdataorcontent,includingtext,imagesorvideos,

basedonuserpromptsandbylearningfromexisting

datapatterns.

Machinelearning:Asub?eldofAI,broadlyde?ned

asthecapabilityofamachinetoimitateintelligent

humanbehaviourandlearnfromdatawithoutbeing

explicitlyprogrammed.4

NLP:A?eldofmachinelearninginwhichmachines

learntounderstandnaturallanguageasspokenand

writtenbyhumans,insteadofthedataandnumbers

normallyusedtoprogramcomputers.

PredictiveAI:AtypeofAIthatusesstatistical

analysisandmachinelearningalgorithmstomake

predictionsaboutpotentialfutureoutcomes,identify

causationandassessrisks.5

Compute:Computereferstotheprocessof

performingcalculationsorcomputationsrequired

foraspeci?ctask,suchastraininganAImodel.It

alsoencompassesthehardwarecomponents,like

chips,thatcarryoutthesecalculations,aswellasthe

integratedsystemsofhardwareandsoftwareusedto

performcomputingtasks.2

Remotesensing:Acquiringinformationfroma

distanceviaremotesensorsonsatellites,aircrafts

anddronesthatdetectandrecordre?ectedor

emittedenergy.AllobjectsonEarthre?ect,absorb

ortransmitenergy,withtheamountvaryingby

wavelength.Researcherscanusethisinformationto

identifydi?erentEarthfeaturesaswellasdi?erent

rockandmineraltypes.6

1DefinitionbytheInternationalTelecommunicationUnion(ITU).

2AINowInstitute.(2023).ComputationalPowerandAI.

3DefinitiontakenfromMicrosoftAzure’sdictionaryoncloudcomputing.

4DefinitionbytheMITSloanSchoolofManagement,basedonthedefinitionbyAIpioneerArthurSamuel.

5DefinitionfromtheCarnegieCouncilforEthicsinInternationalAffairs.

6DefinitionbyNASAEarthdata.

AIforAfrica:Usecasesdeliveringimpact

3/76

Executivesummary

ThepotentialofAIinAfrica

AIholdsimmensepotentialtoboostAfrica’s

economyandtosupporttheSustainable

DevelopmentGoals(SDGs)onthecontinent.WhileAI

isalreadybeingdevelopedanddeployedtosupport

arangeofusecasesacrossAfricancountries,little

researchhasfocusedonbuildingabodyofevidence

ofAIusecasesfordevelopmentonthecontinent.

Thisreportisbasedontheanalysisofover90use

caseapplicationsidenti?edinKenya,Nigeria,and

SouthAfrica–whichbene?tfromthrivingtech

ecosystems–acrossagricultureandfoodsecurity,

energy,andclimate.WhilemanyAIusecasesare

relativelynascent,withsomebeingdeployedas

partofprojectsorpilotschemes,anumberof

commerciallyviablesolutionshavealsoemerged.

Often,AIisbeingincorporatedintoexistingdigital

productsandservices,actingasanenablertomake

digitalsolutionsmorerelevantande?cient,amplify

theirimpact,andfacilitatescaling.

TheagritechsectorisseeingmostoftheAI

innovation,especiallyinKenyaandNigeriawhere

agriculturecontinuestoplayasigni?cantroleinthe

economy.AIisalreadybeingusedforagricultural

advisory,withcompanieslikeTomorrowNowand

ThriveAgricprovidingfarm-levelinsightstofarmers,

andfor?nancialserviceswithcompanieslike

ApolloAgriculturedevelopingalternativecredit

assessmentmethods.AIisalsobeingdeployed

intheenergysector,especiallyinNigeria,where

emergingtechnologieslikeInternetofThings(IoT)

actasanentrypointforadvanceddataanalytics

insmartenergymanagement.Usecasessuchas

energyaccessmonitoringandproductiveuseasset

?nancing,developedbycompanieslikeNithio,

remainatadevelopingornascentstagebutpresent

signi?cantpotentialtoreduceenergypoverty.AI

isalsosupportingclimateusecasesespeciallyfor

biodiversitymonitoringandwildlifeprotection

inKenyaandSouthAfrica,drivenbylargetech

companieslikeMicrosoft’sAIforGoodLabandnon-

pro?torganisationssuchasRainforestConnection.

AIforAfrica:Usecasesdeliveringimpact

4/76

AIfundamentalsandenablingenvironment

Theincreasingavailabilityofdatageneratedby

remotesensingtechnologies,suchason-the-ground

sensors,droneswithhigh-resolutioncameras,and

satellites,hasledtothedevelopmentofmany

AI-drivenusecasesacrosssectors.Analysisof

geospatialandremotesensingdata,poweredby

machinelearning(ML),cansupportawiderange

ofusecasesandactivitiessuchasmonitoringsoil

conditionsfore?ectivecropmanagement,mapping

energyaccessino?-gridareastoinformenergy

planning,andmonitoringclimatechangeimpacts

onecosystems.Despitetheseadvancements,the

availabilityoflocallyrelevantdataremainslimitedin

Africaandposesamajorobstacletodevelopingand

deployingtailoredsolutionsthataddresschallenges

thatareuniquetothecontinent.Inadditionto

barriersinaccessinggovernmentanddomain-

speci?cdata,oneofthemostsigni?cantgapsisin

languagedata.Thescarcityoflocallanguagedata

limitstherelevanceofAI-enabledservicesandposes

asigni?cantbarriertothedevelopmentofgenerative

AIsolutionsthatrelyonlanguagemodels.

InfrastructureandcomputecapacityinAfrica

isgrowing,andcountrieslikeSouthAfricahave

emergedasregionalleaders.Increasinginvestments

indatacentresfromlargetechcompaniesandMobile

NetworkOperators(MNOs)inNigeriaandKenyaare

alsodrivingmomentumintheregion,bringingcritical

storageandcomputingcapacitytothelocallevel.

However,thehighcostsofhardwaresuchasGraphic

ProcessingUnits(GPUs)andcloudcomputingstill

constituteamajorbarriertoAIdeploymentand

adoption,especiallyforlocalentrepreneursand

researcherswithlimited?nancialresources.Aslocal

computeecosystemscontinuetodevelop,thereisan

opportunityforcountriesinAfricatotapintotheir

mobile-?rstmarketstobuildcapacityindistributed-

edgecomputing.InKenyaforexample,deeptech

companyFastaggerdevelopsMLcapabilitiesonedge

devices,includingonlower-endsmartphones.

Acrosscountries,asigni?cantskillsgapstill

underminesthedevelopmentoftheAIecosystem

andusecases.Whileuniversitieso?erAI-related

courses,theyoftenfailtokeeppacewithindustry

needs,andstudentshavelimitedopportunitiesfor

practicallearningandhands-onexperiences.Thereis

alsoadisproportionatefocusoncoreAIskills,suchas

MLanddatascience,withlessemphasisonbuilding

themultidisciplinaryskillsetsneededtoleverage

AItoaddresspressingsocioeconomicchallenges.

Despitethesechallenges,organisationslikeData

ScienceNigeria(DSN)o?erupskillingandmentorship

programmestobuildapipelineofAItalent.In

parallel,endusersrequireafoundationallevelof

digitalliteracytoaccessAI-enabledservices,which

areprimarilyaccessiblethroughdigitalchannelslike

mobiledevices.However,lackofknowledgeand

skillsremainsoneofthegreatestbarrierstoadoption

anduseofdigitaltoolsandservices,especiallyfor

women,low-incomeandruralcommunities,and

personswithdisabilities.

WhileKenya,NigeriaandSouthAfricaareallregional

techleadersandhavesoliddigitalfoundationsthat

canserveasthebuildingblocksforAIdevelopment,

keychallengesremainintheecosystem.Despite

wideenthusiasmaboutthepotentialofAIforAfrica

forexample,privatesectorinvestorsremainrisk-

averseaboutinvestingindeeptech,andstartups

havetorelyongrantfundingfromdevelopment

partnersanddevelopment?nanceinstitutions(DFIs).

Similarly,lowpublicandprivatesectorinvestmentin

ResearchandDevelopment(R&D)mayundermine

thedevelopmentoflocalsolutions.Whilesome

countriesinAfricahavealreadydevelopednationalAI

strategies,Kenya,NigeriaandSouthAfricaarestillin

theprocessofdraftingtheirown–buthaveadopted

inclusiveformulationprocesses.Mostframeworks

acrossthecontinentremainintheirinfancy,

highlightingtheneedtoshiftfrompolicyformulation

toimplementationandtoensureethical,responsible

andsafeuseofAI.

AIforAfrica:Usecasesdeliveringimpact

5/76

High-levelrecommendations

Di?erentstakeholders–governments,development

partners,DFIs,NGOsandCivilSocietyOrganisations

(CSOs),largetechcompaniesandstartups,and

researchandacademicinstitutions–cantakea

numberofactionsandcollaboratetoensurethat

impactfulinnovationsinAfricacanbedeployedand

scaled.Thisinvolvesinvestingindomain-speci?c

andlocallanguagedata,adoptingparticipatory

approachestodatacollection,unlockingaccessto

existingdatasources,andensuringdataprivacyand

security.Strengtheningbaselineinfrastructureand

promotingrenewableenergy,providinghardware

andcloudcredits,enhancingedgecomputing

capabilitiesandbuildinginstitutionalcapacitywillbe

essentialtoboostlocalcomputecapacity.Inaddition,

fosteringacademic-industrycollaboration,raising

awarenessandbuildingcapacityinthepublicsector

willbeessentialtocreateapipelineofAItalentwhile

ensuringinformedpolicymaking.Tofosteradoption

andusageofAI-enabledservices,enhancingdigital

skillsamongendusersandintegratingemergingskills

likeprompt-engineeringintoupskillingprogrammes

willbekey,especiallyasgenerativeAIsolutions

graduallygrowinAfrica.

Stakeholdersacrosssectorscanalsofocuson

supportingthewidertechandAIecosystemto

fosteranenvironmentconducivetoinnovation

andAIdeploymentacrossusecases.Thisinvolves

engaginginpartnershipstounlockaccesstocritical

resourcesforAIentrepreneursandresearchers,and

tosupportthedevelopmentoftheAIecosystem

throughdata-sharingorinfrastructure-sharing

initiatives.Adoptingaconsortium-basedapproach

hasthepotentialtohelpaddressthe?nancinggap,

whileadoptinginnovative?nancemechanisms

cande-riskinvestments.Combiningfundingwith

technicalassistanceandgo-to-marketsupportcan

alsohelpfoundersintheirscalingjourney.Increased

R&Dspendingwillbeessentialtosupportlocal

researchcapacity,whilelocal-globalknowledge

exchangecandrivefurthermomentumandraise

awarenessaboutlocalinnovation.Ascountrieswork

ondevelopingnationalAIstrategies,itwillbecritical

toensureacollaborativeandinclusiveprocess,to

includeprinciplesfortheethicalandsafeuseofAI,

andtoestablishaclearroadmapforimplementation.

Policymakerscanalsoconsiderrollingoutregulations

inaphasedmannertoallowinnovationto?ourish.

AIforAfrica:Usecasesdeliveringimpact

6/76

1.Introduction

AIforAfrica:Usecasesdeliveringimpact

7/76

Overthepastyear,arti?cialintelligence(AI)and

itstransformativepotentialhascapturedglobal

attention.ThepotentialofAIinhelpingachievethe

2030SustainableDevelopmentGoals(SDGs)iswell

established.7,8AIapplicationscancreatesocialand

economicimpact,especiallyinlow-andmiddle-

incomecountrieswhereinnovativeapproachesto

inclusiveandsustainabledevelopmentaremost

needed.Africarepresentsonly2.5%oftheglobalAI

market,yetrecentestimatessuggestthatAIcould

increaseAfrica’seconomyby$2.9trillionby2030—

theequivalentofincreasingannualGrossDomestic

Product(GDP)growthbythreepercent.9Thisboost

ineconomicgrowthcouldtranslateintosigni?cant

developmentimpactsforthecontinent,providing

employmentopportunitiesandhelpingtoraise

millionsoutofpoverty.

Figure1Figure2

EstimatedannualvalueoftheAI

marketinAfricarelativetotheglobal

market

PotentialvalueaddedbyAItothe

Africaneconomy

($trillion,2024-2030)

($trillion,2023)

6

WithAI

5

$2.9trillion

4

$16.5trillion

GlobalAIvalue

3

WithoutAI

$0.4trillion

AfricanAIvalue

Approx2.5%ofthe

globalAImarket

2

2024202620282030

GDPvaluewithAIGDPvaluewithoutAI

CumulativegainsinGDPaddedbyAI

7Smith,M.&Neupane,S.(2018).Artificialintelligenceandhumandevelopment:Towardaresearchagenda.IDRC.

8Bankhwal,M.etal.(2024).AIforsocialgood:Improvinglivesandprotectingtheplanet.McKinseyDigital.

9AI4DAfrica.(2024).AIinAfrica:Thestateandneedsoftheecosystem.

AIforAfrica:Usecasesdeliveringimpact

8/76

MobileconnectivityinSub-SaharanAfricacontinues

todrivedigitaltransformationandsocioeconomic

advancements.Agrowingproportionofthe

populationisconnectedtoandusingmobile

internet,andsmartphonepenetrationisexpectedto

reach88%by2030,creatingnewopportunitiesfor

digitalinclusionandusageofAI-enabledservices.10

CountriessuchasKenya,NigeriaandSouthAfrica

alreadyhavesomeofthemostadvancedtech

ecosystemsintheregion.Kenyaisparticularly

renownedforpioneeringmobilemoneythrough

M-Pesa,whileNigeriahasproducedseveralAfrican

unicorns.Thesecountriesalsohavetech-related

policiesthathavefosteredarelativelyconducive

environmentforinnovationandentrepreneurship.

Theirsoliddigitalfoundationscanserveasbuilding

blocksforthedevelopment,deploymentand

adoptionofAI.

However,unlockingthepotentialofAIwillrequire

overcomingseveralchallenges.Whilethecoverage

gaphassigni?cantlyreduced,theusagegapin

Sub-SaharanAfricastillstandsat59%,meaning

thatmillionsofpeoplewholivewithinthefootprint

ofamobilebroadbandnetworkarenotusing

mobileinternet.Signi?cantdigitaldividesexistand

disproportionatelya?ectlow-incomegroups,those

whoarelesseducated,ruralpopulationsandwomen,

anddigitalisationandAIriskexacerbatingexisting

socioeconomicinequalities.Kenya,Nigeria,andSouth

Africahavecriticalinfrastructuregapsandundergo

regularpoweroutages.Inaddition,insu?cient

availabilityofdataandlackofdataecosystems,low

levelsofdigitalskillsandliteracy,fragmentedornon-

enforcedpoliciesandnascentresearchcapacities

constitutekeybarriersforthedevelopmentoftheAI

ecosystem.AIalsobringssigni?cantrisksintermsof

dataprivacy,biasanddiscriminationthatneedtobe

addressedtoensuresafeandresponsibleuseofthe

technology.

Whiletherehasbeenanaccelerationoftechnology

companiesleveragingAIandinitiativestodevelop

andpromotetheuseofAIonthecontinent,

thesehavenotnecessarilyfocusedonaddressing

socioeconomicordevelopmentchallenges.Most

existingusecasesaretypicallyfoundinsectorssuch

asITservices,computersoftware,ormanagement

consulting.11Thereisalackoffocusonbuildinglocal,

inclusive,andsustainableAIsolutions12thatcan

helpaddresstheSDGsinAfrica.Thereisapressing

needtoidentifyandtestmodelsandusecasesthat

canaddressdevelopmentchallenges,aretailored

tomeetthespeci?cneedsoflocalcommunities,

andhavethepotentialtobescaledtoamplifytheir

impact.Consideringthediversecontextsandcultures

acrossAfrica,fosteringequitablepartnershipsto

buildAIusecasesfordevelopmentandnurturethe

growthoflocalecosystemswillbecriticaltoharness

thepotentialofAItohelpachievetheSDGsonthe

continent.

10GSMA.(2023).TheMobileEconomySub-SaharanAfrica2023.

11TheAIMediaGroupSouthAfrica.StateofAIinAfricaReport2022.

12Inthisreport,local,inclusiveandsustainableAIsolutionsreferstoAIapplicationsthataretailoredtolocalneedsandconstraintstofosterinclusivityandprioritise

addressingdevelopmentchallengesinlinewiththeSDGs.

AIforAfrica:Usecasesdeliveringimpact

9/76

2.Researchobjectivesand

methodology

AIforAfrica:Usecasesdeliveringimpact

10/76

Researchobjectives

ThisresearchseekstoidentifyAI-enabledusecases

andsolutionsthataddressdevelopmentchallenges

relatedtoagricultureandfoodsecurity,energyand

climateaction.ItfocusesonKenya,Nigeriaand

SouthAfrica,whoarealltechnologyleaderson

thecontinentandintheirsub-region,andpresent

signi?cantpotentialtoleverageAIfordevelopment.

Morespeci?cally,theresearchseeksto:

1.IdentifyAI-enabledusecasesandsolutions

acrosstheselectedsectors,highlighttheirkey

requirementsandassesstheirpotentialforimpact,

scaleandconstraints.

2.ProvidealandscapeoverviewoftheAIecosystem

ineachcountrytoidentifygapsandopportunities

toimprovetheenablingenvironmentand

developmentofAI-enabledusecases.

Toaddresstheobjectivesoftheresearch,we

investigatedthefollowingkeypillarsoftheAI

ecosystemtounderstandhowtheyimpactthe

developmentandscalabilityofusecases:the

digitaleconomyfoundations,encompassingdigital

infrastructure,humancapitalandskills,andpolicy

andregulation;theAIfundamentals,includingdata,

AI-speci?cskills,andcomputecapacity;andcross-

cuttingenablers,suchaspartnerships,?nancing

mechanisms,andresearchanddevelopment.

Whilewehaveseparatedtheenablersofeachmain

pillarinourframework(Figure3)toshowthattheAI

ecosystemsitsontopofabroaderdigital/technology

ecosystem,wehavegroupedtheminthereportas

certainelementssuchasinfrastructure,skillsand

policycanbeunderstoodaspartofaspectrum.

3.O?erasetofrecommendationsforkey

stakeholders,pinpointingwaystocatalysethe

developmentoftheAIecosystemdelivering

impactintheregion.

Figure3

TheAIecosystemframework

i

Partnerships

Financing

mechanisms

Data

Researchand

development

AIfundamentals

ComputeAIskills

Digitaleconomyfoundations

Digital

infrastructure

Humancapital

andskills

Policyand

regulation

AIforAfrica:Usecasesdeliveringimpact

11/76

Thisreportreliesondesk-basedresearchandvalidationthroughdiversestakeholderengagement.

ThemethodologyisoutlinedinTable1.

Table1

Researchmethodology

DatasourceObjective

Desk-basedresearch

Reviewincludedgreyliteratureand

industry-speci?creports,academic

publications,databas