HOWTOCOMPETEINTHEAGEOFARTIFICIAL INTELLIGENCE IMPLEMENTINGACOLLABORATIVEHUMAN-MACHINESTRATEGYFORYOUR BUSINESS SoumendraMohantySachinVyas HowtoCompeteintheAgeofArtificialIntelligence:ImplementingaCollaborativeHuman-MachineStrategyforYourBusiness SoumendraMohantyKolkata,India SachinVyasPune,India ISBN-13(pbk):978-1-4842-3807-3/10.1007/978-1-4842-3808-
0 ISBN-13(electronic):978-1-4842-3808-
0 LibraryofCongressControlNumber:2018958917 Copyright©2018bySoumendraMohanty,SachinVyas Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartofthematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation,broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformationstorageandretrieval,electronicadaptation,putersoftware,orbysimilarordissimilarmethodologynowknownorhereafterdeveloped. Trademarkednames,logos,andimagesmayappearinthisbook.Ratherthanuseatrademarksymbolwitheveryurrenceofatrademarkedname,logo,orimageweusethenames,logos,andimagesonlyinaneditorialfashionandtothebenefitofthetrademarkowner,withnointentionofinfringementofthetrademark. Theuseinthispublicationoftradenames,trademarks,servicemarks,andsimilarterms,eveniftheyarenotidentifiedassuch,isnottobetakenasanexpressionofopinionastowhetherornottheyaresubjecttoproprietaryrights. Whiletheadviceandinformationinthisbookarebelievedtobetrueandurateatthedateofpublication,neithertheauthorsnortheeditorsnorthepublishercaneptanylegalresponsibilityforanyerrorsoromissionsthatmaybemade.Thepublishermakesnowarranty,expressorimplied,withrespecttothematerialcontainedherein. ManagingDirector,ApressMediaLLC:WelmoedSpahrAcquisitionsEditor:ShivaRamachandranDevelopmentEditor:LauraBerendsonCoordinatingEditor:RitaFernando CoverdesignedbyeStudioCalamar DistributedtothebooktradeworldwidebySpringerScience+BusinessMediaNewYork,233SpringStreet,6thFloor,NewYork,NY10013.Phone1-800-SPRINGER,fax(201)348-4505,e-mailorders-ny@,orvisit.ApressMedia,LLCisaCaliforniaLLCandthesolemember(owner)isSpringerScience+BusinessMediaFinanceInc(SSBMFinanceInc).SSBMFinanceIncisaDelawarecorporation. Forinformationontranslations,pleasee-mailrights@,orvisit/rights-permissions. Apresstitlesmaybepurchasedinbulkforacademic,corporate,orpromotionaluse.eBookversionsandlicensesarealsoavailableformosttitles.Formoreinformation,referenceourPrintandeBookBulkSaleswebpageat/bulk-sales. Anysourcecodeorothersupplementarymaterialreferencedbytheauthorinthisbookis­availabletoreadersonGitHubviathebook’sproductpage,locatedat/9781484238073.Formoredetailedinformation,pleasevisit/source-code. Printedonacid-freepaper Contents AbouttheAuthors�����������������������������������������������������������������������������������������vAcknowledgments�����������������������������������������������������������������������������������������viiIntroduction���������������������������������������������������������������������������������������������������ix Chapter1:TheEconomicsofArtificialIntelligence������������������������������1Chapter2:ReimaginingCompetitiveAdvantageintheAIWorld�������41Chapter3:BoardtoCEO:“What’sYourAIStrategy?
”�������������������������75Chapter4: InsidetheBlackBox:UnderstandingAIDecision Making�������������������������������������������������������������������������������������91Chapter5:IntelligentProcessAutomation=RPA+AI���������������������125Chapter6:CybersecurityandAI����������������������������������������������������������������143Chapter7:IntelligenceofThings=IoT+Cloud+AI�������������������������155Chapter8:ITOperationsandAI�����������������������������������������������������������173Chapter9: DecentralizedAutonomousOrganizations= Blockchain+AI+IoT�����������������������������������������������������������189Chapter10:EthicsandAI�������������������������������������������������������������������������207Chapter11: PuttingItAllTogether:TowardaHuman-Machine CollaborativeEcosystem�����������������������������������������������������215 Index�������������������������������������������������������������������������������������������������������������231 AbouttheAuthors SoumendraMohantyisanlaimedthoughtleaderandSMEintheareasofanalytics,IoT,AI,cognition,andautomation.Heisatwodecadeveteranwithexpertiseinnext-genBigDatasolutions,BIarchitectures,theenterprisedatawarehouse,customerinsightsolutions,andindustry-specificadvancedanalyticssolutions.Withhisbroadexperience,hehasdesignedandimplementeddataanalyticssolutionsforFortune500clientsacrossindustryverticals.SoumendraisanadvisorwiththeHarvardBusinessReviewAdvisoryCouncil.HeisalsoassociatedwiththeIndianStatisticalInstituteandvariousuniversitiesasavisitingfacultymemberspecializinginBigDataandanalytics.Soumendraspeaksatvariousglobalforums,CAOadvisoryforums,andeducationalinstitutions.HeisauthorofseveralbooksincludingBigDataImperatives(Apress). SachinVyasisanentrepreneurwithover20yearsofexperienceintechnology,data,andanalytics. HewasthefounderandCEOofAugmentIQDataSciences,panyacquiredbyLTI.LTIisaglobaltechnologyconsultinganddigitalpany.AugmentIQfocusedoncreatingplatformsandsolutionsspecifictoBigDataengineeringanddatasciencesandprovidedsolutionstolargeplexdataandanalyticsproblemsforfinancialpanies. HecurrentlyheadsLTI’splatformsandenablingsolutionswithconvergingtechnologiesacrossdevices,data,puting,andartificialintelligence. SachinistherecipientoftheEMCTransformerAwards-2012andtheSKOCHDigitalInclusionAwards-2012inIndia.HehasamechanicalengineeringdegreefromVNIT,Nagpur,India. Acknowledgments Duringthewritingofthisbook,wewerefortunatetobeamidstmanywellwishers,colleagues,andourfamilies,whohavebeenimmenselyhelpfulandsupportivebyplayingseveralroles—co-writers,coaches,anddown-rightfiercecritics.Thisbookcouldnothavebeenwrittenwithouttheirhelp. Wearealsogratefultothemanyclients,analysts,andstrangerswhounknowinglycontributedtothisbookwhentheypatientlylistenedtoourramblings,downrightstupidquestions,andcrazythoughts.Sometimestheycandidlyadmittedwhensomethingdidnotmakesenseandsometimestheyvociferouslyagreedandnudgedustokeepresearchingandwriting. OureffortstowritethisbookwouldhavebeenfutilewithoutthecontinuousguidancefromtheApressteam(Shiva,Rita,andLaura).Youhaveoursincereappreciationforyoursupportthroughoutthisproject. AspecialthankyoutoRitaFernandoKimfromApressforherhardworkonreviewsandforkeepingusonschedule. --- Iamdeeplyindebtedtomylovingfamily—Snigdha,Pratik,Pratyush,andRexie.Withouttheirlove,affection,andsupport,itwouldbeimpossibletosurviveinthecorporatemadness,letalonewritebooks. —SoumendraMohanty IamgratefultoSoum(SoumendraMohanty)fortheopportunitytoworkwithhimonthisbook.Itishisideaandhisbook,andIamreallygladthatIcouldcontributeinsomeways.Therecouldn’thavebeenabetteropportunityfor“myfirst-book”project!
I'dalsoliketothankShweta,mywife,forherpatienceandsupportthroughoutallmyventures,includingthisone. —SachinVyas Introduction Overthelastseveralyears,aswemetwithhundredsofCxOs,seniorexecutives,businessfunctionowners,technologyleaders,andpractitioners,werealizedthattherearefundamentalquestionsthateverybodywastryingtomakesomesenseof: •WhatisAI?
•Whynow?
•Howdoesitimpactmyfunctionorbusinessorevenpanyatlarge?
•WhatshouldIdo?
Thisbookiscollectionofthoughts,assimilatingideas,andviews—somethoughtprovoking,somemundane—addressingseveralaspectsofhowpeteintheageofartificialintelligence.Onceyoustartreadingthebook,youwillrealizethatthechaptersarewritteninablogishmanner,whichispreciselybydesign.Thisistogiveyoufree-flowingthoughtandhelpyougenerateyourownideas. Withtheshortpreambleoutoftheway,let’sgetstarted. TheoriginalAI-poweredWatsonputerthatconqueredthecontestantsontheJeopardytelevisiongameshowwasaboutthesizeofabedroom,with10oddrack-likemachinesformingthefourwalls.Today’sintelligentmachinesareverydifferent.Theynolongerexistsolelywithinbedroomsizedrooms,butarecloudifiedandrunseveral100“instances”ofAIservicesatonce.Theoutputsareconsumedsimultaneouslybyrecipientsallovertheworld,throughmultiplechannels—smartassistants,smartphones,smartdevices,smartappliances,smartvehicles,smartutilities,smartplants,smartfactories,smartbuildings,smarthomes,etc.Theresult?
Wearelivinginaworldof“always-on”intelligence. ThispervasivenessofAIistriggeringanotherinterestingphenomenon—themoreweuseit,thesmarterites.Anythingitlearnsinoneinteractiongetsimmediatelytransferredtotheotherinteractions.AIisnotonelargemonolithicprogram.Itisactuallybinationofdiversesetsofartificialnarrowagents(ANIs),eachspecializedtodoaparticulartaskandwithcapabilitieslikeconversationalinterface,imagerecognition,voicerecognition, xIntroduction text-to-speech,speech-to-text,logic-deduction,naturallanguageparsing,naturallanguagegeneration,knowledgebase,self-learning,andadaptiveengines. Slowlyandsteadily,apictureoftheAIfutureisemerging.TheAIonthehorizonlooksmorelike“AIasaservice”embeddedintoeverything,andalmostinvisible.Acenturyago,wetransformedeverythingtobepoweredbyelectricitytoaugmenthumancapabilities;goingforwardweareatasimilartippingpoint,wherewewill“AI-fy”everythingtoaugmenthumancapabilitiesandinsomecasesintroduceautonomousAItoeliminatehumantasksaltogether.ThankstotechnologieslikeCloud,GPU,BigData,Blockchain,IoT,MLandDL,theprocessof“AI-fying”willesimpleriningdays,transformingeverythingbyinfusingitwithAI.Infact,don’tgetsurprisedifthebusinessplansofpaniesgoingforwardaregoingtotakeXandaddAI. TaketheexampleofGoogle.Everytimewetypeaqueryorclickonalink,weareinganactiveparticipantintrainingtheGoogleAI.StartingfromhowtodatetohowtoconceivetohowtoraisechildrentoDIYscientificexperimentstohowtomanagerelationshipstohowtoprepareaneffectiveresumetowhatarethemostaskedinterviewquestionsandcorrespondinganswers,witheachofthequerieswearefeedingintoGoogle’ssearchengine,wearehelpingGoogleAIrecordourbehavior.Thisincludeswhatwethink,whatwewanttoknow,whenweseekinformation,whyweareseekinginformation,howweusethatinformation,andsoon.Perhapsin10years,Google’smainproductwillbe“MindasaService”andyouwillbeabletorentthemindofaCEO,developer,orscientist. Thisisthepointwherethingsesketchy.Ethicalpoliciesandassociatedeintoplay. WhatispropellingthismassivegrowthofAI?
Threetechnologybreakthroughsactedascatalysts: •Massivelyputation:Thinking,whichistheprocessofreasoningaboutsomething,isamassivelyparallelprocesswherebillionsofneuronsinourbrainactsimultaneously,passingsignalstootherneuronsthroughlayersworks.Thefinaleisjudgment.Wewerehandicappedtoperformmassivelyparallelprocessinguntilthegraphicsprocessingunit(GPU)wasinvented.TheGPUunlockednewpossibilities,whereworks(looselybasedonthewayneuronsworkinourbrain)canfacilitatehundredsofmillionsofconnectionsbetweenthenodes,almostatasub-secondprocessingtime. Introductionxi •BigData:Theintelligencethatweclaimisoursistaughttousovertime.Whenthehumanbrainseessomethingit’sneverseenbefore,ittakestimetodeducewhattheimageis!
ThesameruleappliestoAI.Thankstodigitizationandproliferationofsmartphones,wehaveesstomassiveamountsofreal-worlddata.Ourabilitytocollect,clean,standardize,andstorethisreal-worlddataprovidesuswithanenormoustraininggroundforAI.Theresult?
Wearebeginningtoseeintelligenceinfusedintoalmosteverything,consequentlytransformingeverythingintoa“smart”thing. •Betteralgorithms:Alotgoesoninsideourbraininordertoanalyzeoursurroundings.Weuseheuristicsandmentalmind-mapstoreason.However,itisincrediblyhardforustocodifythethinkingprocess.Ourearlierendeavorstocodifythereasoningprocessresultedinwritinglengthyprograms,mostlyintheformof“IFthenElse…”andtheseprogramswerenotadaptiveenoughtochanges.Deep-learningalgorithmsgiveusawaytogeneratereasoningfromthedataitself,notplexprogramsconsistingofhundredsandthousandsofconditions,butfromthedataitselfchurningoutpatternsandmendations.Wecannowcollectlotsofdataandapplysophisticatedalgorithmstoarriveatpredictions.Theonlydrawbackisthatsomealgorithmsareplexthatwejustcan’tunderstandtheresultsweget. Toillustratetheimpactofthesethreecatalysts,let’sdiscusswhathappenedafterIBM’sputerDeepBluedefeatedthereigningchessgrandmasterGarryKasparovinafamousman-versus-machinematchin1997.Kasparovhadhissuperlativegeniusmindonhisside;DeepBluehadinstantesstoamassivedatabaseofallpreviouschessmovesplayedthousandsofexperts.Kasparovrealizedthatman-plus-machinewouldbefarmorebeneficialthanman-versus-machine.Theconceptofa“centaurplayer”(ahuman/AI)begantoemergewhereAIaugmentsthehuman-chessplayer’sskills.Theresult?
Today,thebestchessplayeris“Intagrand”—ateamofhumansandseveraldifferentchessAIprograms. Interestingly,anotherpatternbegantoemerge.Insteadofdiminishingtheinterestofhumanchessplayers,AI-enabledchessprogramsactuallyinspiredmorepeoplethanevertoplaychess.PerhapstherewassomethingreallymotivatingandrewardingforthehumanchessplayersandtheAIchessprograms,bothturninguptheantetostayonestepaheadoftheother,andconsequentlybothlearningfromeachothertoebetterplayers.Theresult?
TherearemoregrandmastersnowthantherewerewhenDeepBluefirstdefeatedKasparov. xiiIntroduction Thisbringsupanotherpoint.IfAIcanhelphumansebetterchessplayers,byallmeansitisalsopossibleforAItohelpusebetterinotherspheresoflifeandprofessions. Fantasticsofar!
Howdoallofthesetechnologyadvancesrelatetobusiness?
Well,thesignsarethere,howeversubtleatthispointintimetheymaybe.Theworldaroundus,andhenceourbusinessesandtherolesweplayinabusinesscontext,areincreasinglyimpactedbyAIprevalence.ItwillrequireCxOstoadoptanewapproachtoroledelegation—fromimplementingstrategicAIadvisorstoaugmentingemployees. anizationsvalueempoweredAIasmuchastheyvalueempoweredpeople.Forexample,CEOsmustmakeitclearwhensmartalgorithms,ratherthanhumanassociates,aretobeconsulted.Thiscanbedifficult.SomeofthemostimportantdecisionsregardingmachinelearningareusuallyabouttheextentofauthoritytheAIagentsshouldhave.BusinessleaderswhoautomateafactorynowrecoilattheideaoflettingAIdeterminetheirentirebusinessstrategy.Intheabsenceofcleardelineationofauthorityaswellasountability,human-AIconflictwillresultfromdualempowerment. BusinessmittedtotakingadvantageofAIshouldconsiderthefollowingfourAI-relatedimperatives. •AIasstrategicadvisors:AIcaneffectivelyplaytheroleofthestrategicadvisor.TheprimarydeliverablesoftheAIstrategistaretoreframeproblems,assistinstrategydevelopment,andassistindefiningthecourseofactionandtheexecutionplans.Thesealgorithmswillconstantlyproducedata-driveninsightsandmendoptimalcoursesofaction.Humaninterventionandoversightwillsolelyberequiredtoidentifywhichdecisionsaredeferredtothealgorithmsandhowthedecisionsareimplemented. TheAIstrategistwillhaveabroaderviewacrossthesystemsandbusinessprocessespetitiveintelligence,whichwillinvariablypresentmajoroperationalchallenges,suchasinter-processandinterpersonalconflicts.Hence,itisallthemoreimportantthattheoutputoftheAIstrategistshouldbetransparentandexplainabletoall.Thatwaythetrustfactorwillevolvebetweenhumanandmachine. Introductionxiii •AIastaskexecutioners:TheAIstrategisthasdoneitsjobindefiningtheproblemstatementandexecutionplans,andthehumanassociatehasvalidateditandgivenago-ahead.Whatnext?
Youwillneedalgorithmstoanalyzebusinessprocesses,createcleartaskdescriptionsandgoals,anddefinedetailedservice-levelagreementsandkeyperformanceindicators.Managersandemployeesresponsiblefortaskexecutionwillnowperformhigherordertasks—insteadoffocusingonoperationalizing(allocatingresources,managingemployees,andmanagingescalations)theprojectexecution,theywillspendmostoftheirtimereviewingwhetherthealgorithmsareofferingimprovedesandinnovation.Amongtheimportantbenefitsthatalgorithmsofferarehigh-levelreliabilityandpredictability. •AIasavirtualassistanttoemployees:Eventhemosttalentedemployeeshavetheirlimits.Comparedtotheseemployees,AIalgorithmsaregeniusesboxedintoatask’sscope.Therefore,thequestiononbusinessleaders’mindsiswhethertheaveragemanagerandemployeecaneffectivelyworktogetherwithintelligentagents. ModernenterpriseslikeGoogle,Alibaba,Amazon,Netflix,etc.havealreadystartedusingAIassistantsincaseswhereactionableinsightscanimprovetheproductivityoftheiremployeesandcanhelpachievethebusinesses.Employeeshavenootherchoicebuttolearntotreatthealgorithmsasvalued.Ontheonehand,thecollaborationcultureofthehuman-machinewillmeanmuchmorescalabilityandpredictabilitytothees.Atthesametime,itmaytoalargeextentweakenountabilityinanization.Onereasonforthismaybethatinrapidlyevolvingscenarios,whereeverybody(humanandmachine)isrunningagainsttimetodeliverthees,itmaynotbecleartomanagerswhethertheyneedtoretraintheemployeeorthealgorithms. •AIasananization:Isthispossible?
Well,ifyouhavedonethepreviousthree,thenwhynot?
ItisalreadyhappeningwithmanyWallStreethedgefunds.paniesareallowingAIfullautonomyinsteeringanizationtonewlevelsofrisk,profitability,andinnovation.Executivesinthesehedgefundshavehandedoveranunimaginableportionofthedecision-makingprocesstotheiralgorithms. xivIntroduction FinalThoughts ThefourimperativesofincorporatingAIintotheenterprisemayseemfarfetchednow,butthisprocessisinevitable.AsthecapabilitiesofAIcontinuetoadvance,increasedoversightwillleadtoadditionalinsight,andthesoftwarewilllearncontinuouslyandwillbeviewedasanountableagentratherthaninanimatecode.However,CEOs,boardmembers,andseniorexecutiveswillhaveacriticalroletoplay—theymustcloselymonitorthealgorithmsandpromotesimulationstodeterminetheboundariesofthetechnology.Furthermore,businessleadersshouldbecarefulaboutdoingtoomanyAI-ledactivitiestoosoon,asdoingsomaycreateunforeseenimplicationsonresponsibilityandountability.Theymustensureclarityindeference,delegation,anddirection. Inthisbook,chapterbychapterwetouchuponmanyaspectsofAIinthecontextofanization’sprocesses,strategies,advantages,andconsequences,aswellaswhatitmeansforanindividualinanyrole.InZen,“koans”arejusttoolsorbuildingblockstogettotheendgoal.Similarly,wesincerelyhopethatthisbookwillserveasatooltoguideyouandfindsolutionstotheproblemsyouarefacing. Nowit’stimetogetontothefirstchapterofthebookandlearnabouttheeconomicsofAI,includingwhatdefinesAIandwhatmakesitsadoptionsodifficult!