AI+DPI

ArtificialIntelligence:

ThenextfrontierinDigital

PublicInfrastructure

September2024

AIDPIArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure

RomalShetty

ChiefExecutiveOfficerDeloitteSouthAsia

Bharat’sjourneyinbuildingDigitalPublicInfrastructurehassetaglobalbenchmark.DPIshave

fundamentallychangedhowwe,ascitizens,accessessentialservices,engagewiththegovernment,andparticipateinthedigitaleconomy.ProjectionssuggestthatDPIscouldcontributeupto4%toIndia’s

GDPby2030;it’sclearjusthowindispensablethisinfrastructureistoourfuture.

TheworldisincreasinglyturningtoBharatasaleaderinthisspace.We’reseeinggroundbreakinginnovationsemergeregularly—liketheRBI’smostrecentannouncementoftheUnifiedLending

Interface.Bharat'sgrowingroleasaglobalhubfortransformativeandreplicabledigitalsolutionsisundeniable.

DPIisn’tjustabouttech;it’saboutbridginggapsinfinancialinclusionandresilience.Itdirectlysupportsvulnerablepopulationsinenhancingtheiraccesstoessentialservices.Itcanbeusedtocoordinate

large-scaleeffortstoreducecarbonemissionsorprovideemergencyhealthservices,ultimatelyleadingtoamoresustainableandequitablefuture.Thepossibilitiesareendless.Thisiswhy,atDeloitte,we’vemadeitourmissiontoworkinthisdirection.

Lastyear,wetookanotablestepbylaunchingthefirstDPIplaybookatArohana,Deloitte’sGovernmentSummit.WeaimedtobreakDPIdowntoitscorecomponents,providingnationswithadiagnostic

frameworktoassesstheirstrengthsanddevelopmentareas,helpingthemforgetheirownpath.

Lookingahead,theintegrationofArtificialIntelligencewithDPIpresentsunprecedentedopportunities.AIcouldrevolutioniseservicedeliverybyenablingdata-drivendecision-makingandreal-time

optimisation.Itcanalsosolvelanguagelocalisationchallenges,makingDPImoreinclusive.However,withtheseopportunitiescomechallenges—ensuringdataprivacy,managingrisks,andmaintaininginclusivitywillbecrucialtorealisingAI’sfullpotentialwithinDPI.

ThisPointofViewbyDeloittetitled“AI+DPI.ArtificialIntelligence:TheNextFrontierinDigitalPublic

Infrastructure"setsthestageforexploringhowAIcanbeintegratedwithDPI,creatingafuturewheretechnologydoesn’tjustconnectus,butempowersus.Withthispublication,wehopetosimplify

concepts,shareinsights,addresschallenges,andinspirenewideasthatwillhelpshapeadigital

landscapepoweredbyAI+DPIwhereeverycitizencanthrive.We’reexcitedtostartanewconversationwithourreadersaboutthefutureofAI+DPI.

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure

NSNMurty

Partner

GovernmentConsultingLeader

Wearelivinginanextraordinarytimeofdigitaltransformation,andIndia’sjourneyinbuildingDigitalPublicInfrastructure(DPI)hasemergedasashiningexamplefortheworld.Theseeffortshavemadeservicedeliverymoreefficientandlaidasolidfoundationforaninclusiveandresilientdigitaleconomy.Aswelooktothefuture,it'sevidentthattorealisethefullpotentialofDPI,thenextlogicalstepisto

harnessthepowerofAI.

AIrepresentsanamazingopportunitytoreimaginehowDPIscanimproveefficiency,personalise

servicesandexpandtheirreachtoeventhemostremoteareas.However,itismorethanjustscalingexistingcapabilities;it’saboutimaginingwhat’spossible.

Forcitizens,thismeansaccesstoessentialservicestailoredtoindividualneedsanddeliveredwhenandwheretheyareneededmost.Pictureahealthcaresystemthatcananticipateoutbreaksbeforetheyoccuroraneducationalplatformthatadaptstoeachchild’slearningpace,ensuringthatqualityeducationiswithinreachforeveryone.

Forpolicymanagers,AI-basedreal-timedataanalyticscaninformpolicydecisions,classifyandmakesenseofhugevolumesofunstructured(brokendataacrossthecountry)andincreasetransparency,leadingtoamoreresponsivegovernment.

ThefusionofAIandDPIopensmanynewopportunitiesfortheindustryecosystem.Businessescaninnovateusingthisinfrastructure,creatingproductsandservicesthataddressthespecificneedsofadigitallyconnectedpopulation.TheinsightsgeneratedfromAI-drivensystemscanhelpbusinessesbetterunderstandandservetheircustomers.

AtDeloitte,wearecommittedtoleadingthistransformation.Ourlatestrelease,“AI+DPI:Artificial

Intelligence–TheNextFrontierinDigitalPublicInfrastructure,”reflectsthiscommitment.Wehopethispublicationsparksconversationsandinspirespolicymakers,businessleadersandcitizensaliketothinkdifferentlyabouthowAIcanshapeourshareddigitalfuture.Together,let’sworktowardsaworldwheretechnologyserveshumanityinwaysweareonlybeginningtoexplore.

Contents

1.Purposeandreadership01

2.UnderstandingAI04

3.UnbundlingAI07

4.ArecapofDPI13

5.WhydoIneedAIformyDPI?16

6.ImaginewithAI+DPI19

7.HowAIcomponentsinteractwithDPIbuildingblocks23

8.ManagingAI+DPI27

9.AIasDPG32

10.Notablecasestudies34

11.CalltoAction37

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure

vi

01

Purposeand

readership

AIDPIArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure

02

ThePurpose

Hithere,gentlereaders!Iam

SensAI(hopeyouenjoyedtheplayofwords!).Iamgoingtobeyourguideinthisjourneywherewe

breakdownAIandDigitalPublicInfrastructure(DPI)intotheir

simplestparts.Together,wewillexplorehowthesecomponentscanbeusedtobuildinnovativeproductsforgovernmentservicedeliveryandhowAIcanamplifythepowerofDPItomakepublicservicesmoreefficient,accessible,andimpactful.

AtDeloitte,webelievethatsupportingnation-buildingbeginswithempoweringgovernmentstodeliverdirect,measurable,andscalablebenefitstotheirpeople.Technologyplaysahuge

roleinthis.Whentechisinteroperableandhasthepotentialtocreatenetworkeffects,the

impactcanbeevenmoresignificant.DigitalPublicInfrastructureinIndiahasrevolutionised

governancebyconnectingbillionstoessentialservices,drivingfinancialinclusion,and

unlockingnewgrowthopportunities.ThisiswhyweadvocateforDPIasacriticaltechnology.DeloittelaunchedthefirstDPIplaybookfornationsatthefirsteditionofArohana,Deloitte’s

GovernmentSummit,in2023asaresourcedesignedtohelpcountriesunderstand,implement,andharnessthepotentialofDPIstoacceleratetheirdigitaltransformationjourneys.Our

objectivewastobreakdownDPIasatechnologytoitscoreblockandgivenationsadiagnosticframeworktoassesstheirstrengthsanddevelopmentareasandforgetheirownpath.

We’renowtakingitastepfurtherbylearningaboutAI-poweredDPIswiththefollowingobjectives:

ExplorewhatanAI-integratedDigitalPublicInfrastructuremightlooklikeinpractice.

?HighlighttheadvantagesofcombiningAIwithDPIinimprovingpublicservicedeliveryandefficiency.

?DiscusspotentialrisksandchallengesassociatedwithintegratingAIintoDPI,ensuringinformeddecision-making.

?InspirenewideasandusecasesinthepublicsectorusingAI+DPI.

?Discusspossiblerolesandresponsibilitieseachstakeholdercanplayforthisinnovation.

Hi,IamCurioandI’mheretolearnwithyou!

Whoisthismeanttoreach?

Government

Social

entrepreneursandstartups

Technologyandserviceproviders

Donorsand

fundingagencies

Curiouscitizens

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure03

04

02

UnderstandingAI

ArtificialIntelligencesimplified

Thetermartificialintelligencebroadlyreferstoapplicationsoftechnologytoperformtasksthatresemblehumancognitivefunctionandisgenerallydefinedas“thecapabilityofamachinetoimitateintelligenthuman

behaviour.AIsystemsgenerallyworkbyingestinglargeamountsoflabelledtrainingdata,analysingthatdataforcorrelationsandpatterns,andusingthesepatternstomakepredictionsaboutfuturestates.

AIworksbylearningfromlotsofinformation,justlikehowwelearnnew

thingsatschool.Forexample,ifyoushowanAIthousandsofpicturesof

differenttypesoffood,itwillstarttorecognisewhateachtypeoffoodlookslike.Sonexttimeyoushowitapictureofapizza,itwillknowit'sapizza

andwillalsobeabletodifferentiateitfromnoodles.Thisprocessiscalled"machinelearning”.

Oneofthemostcommonisthevirtualassistant(VA)onyourphone.WhenyouaskVA,"What'stheweatherliketoday?"Itunderstandsyourquestion,looksuptheweatherinformation,andtellsyoutheanswertailoredtoyourpreciselocation.Itfeelsalmostliketalkingtoahelpfulfriend,butit'sactuallyAIatwork.

AIisalsobehindtherecommendationsyouseeoncontentstreaming

websites.Whenyouwatchamovieorbuyaproduct,AIlearnsaboutyourpreferencesandsuggestsotherthingsyoumightlike.It'slikehavinga

personalshopperwhoknowsyourtastesverywell.

/rules-guidance/key-topics/fintech/report/artificial-intelligence-in-the-securities-industry/overview-of-ai-tech

/searchenterpriseai/definition/AI-Artificial-Intelligence

Let’slookatanoverviewofareaswithinAI

Imageanalysis

Knowledgepresentation

Natural

languagegeneration

Speechanalysis

Visualisation

Natural

LanguageProcessing

Graphanalysis

VirtualPersonalAssistance

Q&A

systems

Machine

translation

Recommendationsystems

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure05

06

What’sthebuzzaroundGenerativeAI?

"Iamacyborg,andIcanreadminds!You'reprobably

thinking,‘Alright,AIiscool,butwhat'sallthisbuzz

aboutGenAI?'HowisitdifferentfromtraditionalAI?

WhatimplicationscouldGenAIhaveonDPI?Andmostimportantly,isthereanythingspecificIshouldbeawareofwhenitcomestoGenAI'sintegrationwithDPI?“So

let’sjumprightintoit..

AI

AIistypicallydesignedtorecognisepatterns,makepredictions,or

automatespecifictasksbasedonpre-definedrules.

Typically,outputsdecisions,

ETHICALCONSIDERATIONS

IMPLICATIONSOFGENAI

02

01

03

Dataprivacy

AccountabilityAlgorithmicBias

classifications,orpredictionsOftentrainedonlabeleddata

GenAI

GenAIisasubsetofAIthatcancreatenewcontent,whetherit'stext,

images,audio,code,orevencomplexsimulations,bylearningpatternsfromexistingdata.

Outputsoriginalcontent

Trainedonlargedatasetsto

understandandreplicatethepattern

Itcanbeusedtocreatepersonalisedcontentandservices,improvingtheoveralluserexperienceofDPIsystems.

Itcanrapidlyscaleservicestomeetgrowingdemandwithoutaproportionalincreaseinhumanresources.

Itcanautomatecontentcreationforgovernmentportals,educationalresources,orcommunicationmaterials.

ItcandriveinnovationwithinDPIbyenablingthecreationofnew,AI-drivenservices.AIcouldsimulatedifferentpolicyoutcomes,helpingpolicymakersmakemore

informeddecisions.

/content/dam/Deloitte/nl/Documents/deloitte-analytics/deloitte-nl-data-analytics-artificial-intelligence-whitepaper-eng.pdf

AIDPIArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure07

03

UnbundlingAI

08

CorecomponentsofArtificialIntelligence

Representationandreasoning:Machinereadablestructures,logicandinferencefromfacts,probabilisticreasoningforuncertaininformation

Knowledgeprocessing

Understandingandinteraction

Imagegeneration

Modelevaluation&optimisation

Noisereductionandenhancement

Languagegeneration

Motionanalysis

Reinforcementlearning

Sub-sets

Speakeridentificationandverification

Semantics

Objectdetectionandrecognition

Unsupervisedlearning

Speechsynthesis

Textprocessing

Imageprocessing

Supervisedlearning

Speechrecognition

Naturallanguageprocessing

Computervision

Machinelearning*

Core

componentsofAI

Speechcomprehensionandsynthesismodels

*Althoughacomponent,machinelearningcancontributetowardsenhancementofseveralotherAIcomponentsincludingNLP,imageanalysesetc.

NLPinpublicservicedelivery:Apracticalbreakdown

Let'sbreakdownNLPandunderstandwhatitisinsimpleterms,exploreitsabilitiesingovernance,andseesome

practicalusecasesinpublicservicedelivery….

NaturalLanguageProcessing(NLP)isatechnologythathelpscomputersunderstandandusehumanlanguage.Itenables

machinestoread,listen,andcommunicatewithpeopleinawaythatfeelsnatural.

BroadpracticalapplicationsofNLP

Process,analyse,classifyandarchivelargedocuments

Runchatbotsforautomatedpublicservicedelivery

Analysecitizenfeedbackorcallcenterrecordings

Answerwho-what-when-wherequestions

Languagegeneration

Text

processing

Semantics

Understandingandinteraction

Input

Citizenquery-“Canyouhelpmeapplyforunemploymentbenefits?"

Socialmediapostsregardingpublichealthcampaigns

liketweets,SocialMediacomments

Transcriptsofagovernmentmeeting

Scannedlandrecordsandpropertydeeds

Output

Automatedresponse-“Sure,pleasetellmesincewhen

haveyoubeenwithoutajob"

Analysisreport-"Positivesentiment:65%,Negative

sentiment:20%,Neutral

sentiment:15%"

Summaryreporthighlightingkeypointsanddecisions

madeduringthemeetings

Digitisedandorganised

recordsaccessiblethroughanonlinegovernmentportal

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure09

10

Computervisioninpublicservicedelivery

Unpackingcomputervision!We'lldemystifyitwitheasyexplanations,uncoveritspowerfulusesingovernance,and

showcasereal-worldexamples.Readytoexplore?

Computervisionletsmachinesseeandunderstandtheworld,transformingthemintosmarthelpersthatkeepour

citiessafe,trafficsmooth,andpublicspacesclean.It’slikegivingtechnologytheeyesandbrainstomakelifeeasierand

saferforeveryone!

Broadpracticalapplicationsofcomputervision

Activitydetectionandsurveillance

Biometricandfacialrecognition

Patternrecognitionandanomalydetection

Geospatialanalysisandlandmapping

Imageprocessing

Motionanalysis

Imagegeneration

Objectdetectionandrecognition

Input

Oldphysicalcopyofagovernmentdocument

Surveillancecamerasforacitypark

Trafficlightsonaroute

Planforanewcommunityhealthcentre

Output

Enhancedqualityofscanneddocumentsforbetter

machinereadabilityanddigitalarchiving

Trackobjectsinthepark,

identifyunattendedbagsandalerttherelevantauthorities

Identifypeakpedestrianandvehicletimingstosuggest

peaktimingsandappropriatesignaldurations

3Drenderingsofthe

proposedcommunityhealthcentre,includinginterior

andexteriorviewsbasedondefinedrequirements

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure11

Speechcomprehensionandsynthesisinpublicservicedelivery

Let,sbreakdownspeechcomprehensionandsynthesisintopracticalelements

Speechcomprehensionisliketeachingacomputertounderstandwhatpeoplesay,whilespeechsynthesisisliketeaching

ittotalkback.Imaginehavingasmartassistantthatcanlistentoyourcommandsandrespondwithhelpfulinformation,

justlikeahuman!

Broadpracticalapplicationsofspeechcomprehensionandsynthesis

Interactivevoiceresponse

Contentdelivery

Accessibilitysolutions

Trainingandeducation

Noisereductionandenhancement

Speechrecognition

Speechsynthesis

Speakeridentificationandverification

Input

In-personcommunityforumdiscussinglocalissues

Accessinggovernmentservicesonline

Voicesampleofcitizens

applyingforsocialbenefits

Emergencyresponsecall

Output

Providingaccuraterecordsofthediscussionforpublicrecordsallowsgovernmentofficialstoreviewcitizen

concernsandsuggestionsindetail.

Assistingcitizens,

especiallythosewithvisual

impairments,innavigating

onlinegovernmentservicesthroughclearaudioguidance.

Duringremoteinteractionswithcitizens,VoiceAssess

canverifytheauthenticityoftheindividual,ensuringthatpublicfundsaredisbursedonlytothosewhoaretrulyeligible.

Improvingtheclarityandintelligibilityofemergencycommunications,ensuringthatcriticalinformationisaccuratelycaptured.

12

MachineLearninginpublicservicedelivery

DecodingMachineLearning!Let’sbreakitdown,exploreitswide-ranginguses,anddiveintoreal-worldexamples.

Machinelearningisliketeachingacomputertolearnfromexperience,justlikehumansdo.Itusesdatatorecognisepatterns,

makepredictions,andimproveovertimewithoutbeingexplicitlyprogrammed.Imagineasmartassistantthatgetsbetterat

helpingyouthemoreyouuseit!

Broadpracticalapplicationsofmachinelearning

Predictiveanalyticsandforecasting

Personalisationand

recommendationsystem

Processoptimisationandworkflowautomation

Clusteringandclassification

ofdata

Supervisedlearning

Unsupervisedlearning

Reinforcementlearning

Modelevaluationandoptimisation

Input

Historicalcrimedatawithlabeledoutcomes(e.g.,typeofcrime,location,time)

Surveyresponsesand

feedbackfromcitizensonvariouspublicservices

Routesandschedulesofwastecollectiontrucks

Userinteractiondatawithagovernmentservicechatbot

Output

Predictwherecrimesarelikelytooccurinthefuture,helpinglawenforcementallocateresourcesmore

effectively

Identifiespatternsand

clustersinthedatawithoutpredefinedlabels,helpinggovernmentagencies

understandcommunityneedsandtailorservicesaccordingly

Optimiseroutesforwaste

collectionbybalancing

factorslikefuelconsumption,traveltime,andcoverage,

ultimatelyleadingtomoreefficientwastemanagement

Evaluatingthechatbot's

performancehelpsidentifyareaswhereitcanbe

improved,suchasbetter

understandinguserqueriesorprovidingmoreaccurateinformation

04

ARecapofDPI

AIDPIArtificialIntelligence:ThefrontierinDigitalPublicInfrastructure

LookingbackattheDPIplaybookforNations

DPIPlaybookforNations

,publishedbyDeloittein2023,isacomprehensiveresourcedesignedtohelpcountriesunderstand,implement,and

harnessthepotentialofDPIstoacceleratedigitaltransformationwithinclusiveandsustainableeconomicdevelopment.ThisplaybookhighlightstherolethatDPIinfrastructuresplayinenablinginteroperability,scalability,andgrowthacrosssectors.

Deliverymechanismfordigitalservices

Informationexchange

Sector-specificbuildingblocks

Foundationalbuildingblocks

Physicalinfrastructure

Servicedeliverysystemsacrosssectors

DatasharingAPIs,gateways,secureinformationexchange

LMS,healthsurveillance,

e-marketplace,doctorverification,etc.DigitalIDsandregistries,datasharingandtrustinfrastructure,discoveryandfulfillment,andpayments

Hostinginfrastructure,connectivity,servers,datacentres,etc.

Thisarchitecturecanbeusedbyanyinstitutionacross

sectorstobuildnewserviceswithouthavingtoredesignthe

basicinfrastructure

HowtoimagineDigitalPublicInfrastructure?

TheUserishere

AppsLayer

Informationsolutionstodifferentverticals,e-commerce,cashtransfers,remote

education,telehealth,etc.

Consent

Theconnecting,sectoragnostics

intermediatelaterwhichcontainthefollowingblocksandmore

PhysicalLayer

Thephysicallayerincludesconnectivity,devices,

servers,datacenters,routers,etc.

Payment

Infrastructure

IdentityRegistries

DPILayer

Networks

etc.

DataExchange,

DPIDesignPrinciples

ReusableIntegrableCostEffectiveCustomisable

https://www2deloittecom/content/dam/Deloitte/in/Documents/publicsector/inpsdpiplaybookbrandnoexppdf

14

FoundationalDPIBuildingBlocks

IDsandelectronicregistries

Digitalpayments

Credentials&DataSharing

Discovery&

Fulfillment

DigitalSignature

Interoperable

Scalable

DerivinginspirationfromcountrieswhohavedoneandexperiencedDPIs

India’sDPIstory

India’sDPIjourneyhastakenmassiveleapsinthelast

decade

2.5billion

Vaccinesdeliveredwith

digitallyverifiablecertificates

throughCoWIN

US$322billion

TransferredthroughG2Pinfrastructure

99%

PeopleinIndiahaveadigital

identitynumberintheform

ofAadhaar

~6billion

SecuredigitalcredentialsthroughDigiLocker

Education

Skilling

TheDigitalInfrastructureofKnowledgeSharing(DIKSHA)builtusingSunbird,isafree-to-useschoolplatform

withmultiplesolutionsforstudents,teachers,and

administrators.DIKSHAoffersover8,900coursesand

200,000piecesofcontentacross30Indianlanguagesfrom

11,500contributors,reachingapproximately180millionstudentsand7millionteachers.

SincetheunderlyingtechnologyforDIKSHAisavailableasbuildingblocks,itisbeingusednotjustinprimaryeducationbutalsoinskilldevelopmentforCOVID-19

trainingfordoctors,nurses,andotherhealthworkers.

Health

Finance

IndiabuiltitsvaccinedistributionandmanagementplatformCoWin,builtusingDIVOC,toscaleupvaccinedeliveryusingDIVOC,anopen-sourcesoftwarefordigitalcertification.

DIVOCwasalsodeployedinfourothercountries(Indonesia,Philippines,SriLanka,andJamaica)tohelpfacilitatetheir

vaccinationprogrammes.

Digitalpaymentsystemsforinter-bankpeer-to-peerandperson-to-merchanttransactions.UPIcurrentlyconnectsmorethan50millionmerchantsaspartofitsnetwork.

OtherareasofDPIintervention

ONDC:OpennetworkfordigitalcommercebuiltusingBecknprotocol

ONEST:Opennetwork

foreducationandskilling

transactionsbuiltusing

Becknprotocol

BHASHINI:AI-ledlanguage

translationplatformbuilt

usingSunbirdandAI4Bharat

NationalurbanstackbuiltusingDIGIT

Accountaggregator

frameworkbuiltonDEPA

toenablesecuretransfer

offinancialdata

/view/journals/001/2023/078/article-A001-en.xml#A001fig06

/

/

.in/

https://becknprotocol.io/

/content/dam/Deloitte/in/Documents/public-sector/in-ps-dpi-playbook-brand-noexp.pdf

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure15

05

WhydoIneedAIformyDPI

AI+DPI|ArtificialIntelligence:ThenextfrontierinDigitalPublicInfrastructure17

Fromfoundationtofuture:howAIcanenhanceourDigitalPublicInfrastructure

Theworldismovingtowardsanexperiencewhereeveryinteractionisseamless,whetherit’saccessingessentialservicesormakingfinancialtransactions.ThisworldispoweredbyDPI,thebackbonethatsupportsourdailydigitallives.

Butasthedemandforthisinfrastructuregrows,sodothechallenges.Thesystemispowerful,yetitisboundtostruggleundertheweightofever-increasingdataandcomplexity.Itislike

asophisticatedenginerunningatfullcapacitybutwithouttheabilitytoforeseeproblems,adapttonewdemands,oroptimiseitsperformanceinrealtime.

WhypairAIwith

DPI?Teachme

SensAI!

Now,imagineaddingalayerofArtificialIntelligencetothissystem.

Detect

inefficiencies

ooooo↓

√

√

Data

intelligence

Personaliseservices

Anticipateneeds

Risk

readiness

Learnstoimprove

√

WhenshouldDPIandAI

joinforces?

WithAI,decisionsarenolongerdelayedbythelimitsofhumananalysis—they’remadeinthemoment,informedbyvastamountsofdataprocessedatlightningspeed.Servicesaren’tjustprovided—they’repersonalised,adaptingtotheuniqueneedsofeach

individual.Risksaren’tjustmanaged—they’reanticipatedandmitigatedbeforetheycancauseharm.Theinfrastructureisready,butAIwillallowittoreachitsfullpotential.

18

Thescaleofservicedelivery

Whenservicesneedtobe

deliveredtoalargepopulation

acrossmultipleregions

Interoperabilityrequirements

Y

Whenmultiplesystemsand

stakeholdersneedtocommunicate

andintegrateseamlessly.

D

P

I

?

Complexityofdecision-making

e

d

Whentasksrequiremoderate

d

I

n

e

creativedifficulty,contextvariability,and

W

h

e

n

o

accuracyneeds,likerecording

d

o

I

n

e

e

d

A

I

?

WhendoI

need

DPI/AI?

Costandresourceefficiency

Whenthere’saneedto

optimiseresourcesand

reducecostsacrosslarge-

scaleoperations.

regulatorycompliance

W

h

e

n

Automationofrepetitive,lowcontexttasks

Whenthere’saneedtoautomatehighly

repetitivetasks,requiringhighaccuracyand

lowcontextvariability

Personalisationanduser

interaction

Whenservicesneedtobe

personalisedforindividualusers

InclusivityandAccessibility

Whentheinfrastructuremustservediverse

populationsintermsoflanguage,literacylevels,

disabilities,agegroups,andsocio-economic

backgrounds.

06

ImaginewithAI+DPI

AIDPIArtificialIntelligence:ThefrontierinDigitalPublicInfrastructure

20

BUS

STOP

WithAI+DPI

WithDPI

BUS

STOP

Thefutureoftrafficmanagement

Scenario

BUS

STOP

Acrowdedbusstopwithlongqueues.

Busesarriveirregularly,andcommuters

lookfrustrated.

Amobileappshowingbuslocationsinreal

time.Passengerscanbooktheirticketsin

advance.Busesarrivemorepunctually,and

queuesstarttoshorten.

IDs&electronicregistries

Discovery&fulfilment

Digitalpayments

AIdynamicallyadjustsbusroutesand

dispatchesextrabusesduringpeakhours.

Commutersboardbuseswithease.Any

bottlenecksandchallengesarepredicted,and

therelevantauthoritiesarealertedinadvance.

NLP

BUS

STOP

WithAI+DPI

WithDPI

BUS

STOP

Thefutureofhealthcare

Scenario

BUS

STOP

Acrowdedclinicwithlongwaitinglines.

Patientslookfrustrateda