final eport #1 for batch 5 of the iot standards

FINALREPORT#1FORBATCH5OFTHEIOTSTANDARDSDEVELOPMENTPROJECTLITERATUREREVIEW:ETHICALISSUESANDSOCIALACCEPTABILITYOFIOTINTHESMARTCITY

FEBRUARY2018

Preparedfor

VilledeMontréal

TotheattentionofMr.Jean-MartinThibault

Director(CTO),ITArchitecture,InnovationandSecurity

VilledeMontréal

275,rueNotre-DameEst

Montréal,QC,HCY1C6

Canada

VilledeMontréal FinalReport

February2018

IoTintheSmartCity:EthicalIssuesandSocialAcceptability Page2

This reportwaspreparedbyCIRAIG (Centre internationalde reference sur le cycledeviedesproduitsprocédésetservices).

CIRAIGwasestablishedin2001toprovidebusinessesandgovernmentwithacademic,state-of-the-artexpertiseonsustainabledevelopmenttools.CIRAIGisoneoftheworld’sleadingcentresoflifecycleexpertise.TheorganizationworkswithmanyresearchcentresthroughouttheworldandactivelyparticipatesinthelifecycleinitiativeoftheUnitedNationsEnvironmentProgramme(UNEP)andtheSocietyofEnvironmentalToxicologyandChemistry(SETAC).

CIRAIGhasdevelopedrecognizedexpertiseinlifecyclestools,includingenvironmentallifecycleassessment (ELCA) and social life cycle assessment (SLCA). Its research complements thisexpertise,withstudiesonlifecyclecostanalyses(LCCAs)andothertools,includingcarbonandwater footprints. CIRAIG’s activities include applied research inmany critical sectors, such asenergy,aerospace,agrifood,wastemanagement,pulpandpaper,minesandmetals,chemicalproducts, telecommunications, finance, urban infrastructuremanagement, transportation andgreenproductdesign.

DISCLAIMER

Theauthors are responsible for the selectionandpresentationof their findings. TheopinionsexpressedinthisdocumentarethoseoftheprojectteamanddonotnecessarilyreflecttheviewsofCIRAIG,PolytechniqueMontréalorESG-UQÀM.WiththeexceptionofdocumentsproducedbyCIRAIG(suchasthisreport),anyuseofthenameofCIRAIG,PolytechniqueMontréalorESG-UQÀMinpublicdisclosuresrelatingtothisreportmustreceive prior written consent from a duly appointed representative of CIRAIG, PolytechniqueMontréalorESG-UQÀM.

CIRAIG

Centreinternationaldereferencesurlecycledeviedesproducts,procédésetservicesPolytechniqueMontréalDépartementdegéniechimique3333,cheminQueen-Mary,suite310Montréal(Québec)CanadaH3V1A2www.ciraig.org


February2018


WorkingGroupResearchTeamExecution

SaraRussoGarrido Supervision,ResearchandWriting

Marie-ClaudeAllard ResearchandWriting

JoanieBéland ResearchandWriting

EmmanuelleCaccamo ResearchandWriting

TylerReigeluth ResearchandWriting

Jean-PhilippeAgaisse ResearchandWriting

Marie-LucArpin Editing

ProjectAdministration Prof.NicolasMerveille,PhD Professor,ESG,UQAMandCIRAIG

ProjectParticipantsfromtheVilledeMontréalJean-Martin Thibault, Pierre-Antoine Ferron, Stéphane Guidoin, Michel Charest, Song NhiNguyenandMartin-GuyRichard.


February2018


Summary

Introduction

Since2014,Montréalhasbeenimplementingastrategicinitiativetobecomeaninternationallyrenownedleaderamongsmartanddigitalcities.Montréalplanstodeveloponitsown—andinpartnershipwithresidents—technologicalsolutionstothemetropolis’skeychallenges,whichitwilldeploytransparently,withadvancedtechnologiesandonahumanscale(VilledeMontréal,2017).

Installationoftechnologicalinfrastructureiscentraltothedeploymentstrategy,whichincludesthe development of the city’s Internet of Things (IoT), accompanied by the creation of atechnologicalandanalyticaldatacollection,storageandanalysissystem.IoTalsomeansinstallingamultitudeofsensorsaroundthecitytogatherawidevarietyofdataonassetsandactivitiesthroughoutthemetropolis.Oneofthisdigitalstrategy’skeyfeaturesisthecollectionofdatafromsensorsandexternalsourcesforinternalusebythecityanditspartners,alongwithexternalusethroughbigdatareleases.

WhiletheexploitationofdataacquiredthroughIoToffersopportunitiestoinnovateandimproveMontreal’squalityoflife,itraisesethicalissuesandrisksthatcouldtriggersocialopposition.

Liketheadventofurbanverticality1inthelate20thcenturyandthecreationofahighwaysysteminthe1930s,IoTrepresentsarebootoftheurbanstructure—onebuiltaroundtheconvergencearisingfromEconomy4.0.2Thistransformationofurbanarchitectureisnotonlytheoutgrowthoftechnologicalevolution,butadrivingforceinsocialchange.

ThisreportoutlinespotentialissuesofethicsandsocialacceptabilitypertainingtourbanIoTandidentifiespotentialsolutions,tosupportthecityinitsdeliberationsonthistopic.Thisdocumentalso serves as a stepping-stone to subsequent project phases, such as finding a conceptualframeworkfordevelopingaprogramtostudy,manageandaddress issuesofethicsandsocialacceptabilitywithrespecttotheIoTproject.

1Theemergenceofhighrises,withpeoplelivingonefloorovertheother,hasgivenrisetounprecedentedpopulationdensities.2The“4.0Economy”emergedoutofthefourthindustrialrevolution.WhilethethirdindustrialrevolutionmadeelectronicsandITcentraltosociety,thefourthischaracterizedbyamergeroftechnologiesthatblurboundariesbetweenthephysical,digitalandbiologicalspheres.Algorithmicanalysis,theInternetofThingsandbigdataarecoretechnologicalcomponentsofthe4.0Economy(Schwab,2017).


February2018


BasicConcepts:EthicalIssuesandSocialAcceptability

Ethicalissuesareidentifiedinthisreportwheneverabasicvalueormoralprinciplecomesintoplay inaparticularmatterorsituation(Commissiondel’éthiqueenscienceettechnologieduQuébec,2017).Asweknow,ethicsisastatementofthecorevaluesandprinciplesthatshouldguideanddirectourinteractionswithothers.Thesevaluesandprinciplesgivemeaningtoourlives and enable us to distinguish, in a given context, between good/evil, right/wrong, andappropriate/inappropriate.3

“Socialacceptability”isacontroversialterm,whichhasbeenthesubjectofnumerousdebatesover its definition (Gendron, 2014; Battelier, 2015). For this project, we shall use Gendron’sdefinitionofsocialacceptability(2014)as“thepublicsentimentthataplanordecisionresultingfromcollectivewisdomisbetterthanknownalternatives,includingthestatusquo.”

IoTTechnologicalandAnalyticalFlowchart

OurreviewoftheliteratureisdesignedtocoveralltopicsrelevanttothespecificcaseofdeployingIoTinMontréal.FigureAisaflowchartofthissystem’stechnologicalandanalyticalcomponents.

3Wehaveadoptedapluralistapproachtoethicsinthisprojectaimedatidentifyingissuesraisedintheliterature,whatever theauthors’ethicalviewpoint.Thisapproach,whichhasalsobeenusedelsewhere(suchas theEU’sETICAproject),permits takingdifferentoutlooksand interpretations intoaccountanddelineatingcoexistingperspectivesinthisfield(Stahl,2011).


February2018


FigureA:IoTTechnologicalandAnalyticalFlowchart

SteponeinvolvesplanningtheurbanIoTinstallation/implementationproject.Next,sensors,aswellashardwareandsoftwareforoperatingthem,willbedeployed.Datawillthenbecollectedfromthesesensors,aswellasfromexternaldatabases,suchassocialmedia,andtransferredtostorage, and eventually archival, sites. Data will be processed (aggregation, possibleanonymization, etc.) at different sites.While these intermediate steps do not appear on theflowchart, they are important. Data analysis (including predictive and prescriptive) is thenanalyzed at different points. These analyses primarily serve two key audiences—municipaldecisionmakersand theirpartners (suchas theSTM), and thepublic, including residentsandbusinesses,byreleasingopendataanddevelopingappsforMontrealersdesignedtoimprovethequalityofurbanlife.Section3.2.1ofthisreportdescribesthissystem’stechnicaldetails.Italsoexplainswhysuchasystemmustevolveinthepresenceofbigdataandnotinavacuum.Oncecollectedandreleased,dataiscombinedwithexistingdatatakenfrominsideandoutsidethecityadministration.

Thissystembreaksdown into fourmainphases in termsof theethical issues identified inourreviewoftheliterature:

1. Infrastructureplanningandmaintenance.2. Datacollectionandstorage(includingprocessing).3. Internal/externaldataanalysis.4. Releaseofopendataandestablishmentofpublic,digitalservices.

.

.

Installa'on)

Transport)

Stockage)temporaire)

Stockage)ins'tu'onnel)

Analyse)temps7réel)

Analyse)de)flux)

Archivage))

Analyses)interac've)&)en)lot)

Ouverture)des)

données)Analyse)(données)

périodiques)ou)interac'ves))

Visualisa'on)des)données)

et/ou)Résultats)pour)prises)de)décisions)

Analyse)et)développement)d’applica'ons)

par)les)citoyens/entreprises))

Applica'ons)offertes)aux)citoyens)

Légende:()))))))))))))))) )Analyse)des)données)) )Installa'on)des)technologies)et)flux)physique)des)données))

Planifica'on)du)projet)d’IdO)

Destruc'on)données)

Données)captées)

Données)externes)


February2018


6

.

.

Enjeux'de'transforma0on

1.  Transforma+on,en,espace,prévisible,et,individualisé,

2.  Impact,sur,l’environnement,

Transforma+on,de,la,ville,

1.  Décisions,pour,l’op+misa+on,et,non,l’op+mum,social,

2.  Dilu+on,de,la,responsabilité,décisionnelle,

3.  Déterminisme,4.  Réduc+on,du,champ,des,

possibles,

Transforma+on,de,la,gouvernance,

Enjeux'éthiques

Inclusion,

Indépendance,des,pouvoirs,publics,

Transparence,et,fiabilité,

Vie,privée,

Liberté,

Poten0els'freins'à'l’acceptabilité'sociale'

SocialAcceptability

WecannotconfidentlyidentifyissuespertainingtothesocialacceptabilityofIoTinthesmartcity,since theexisting literature contains so few studieson this topic.However, Section10of thisreportlistsexistingstudiesfocusingonsocialacceptabilityduringtheusephase,particularlyintermsofthevalueandutilityofproducts,servicesandinfrastructure.We cannot present a clearly defined list of social acceptability issues, since there are so fewstudiesonthetopic.ThisreportaccordinglypresentsethicalissuesandconcernsidentifiedintheliteraturethatcouldbesourcesofpublicresistancetotheurbanIoTprojectandmaritssocialacceptability.Theseethicalissuesandconcernsappearinthefollowingfigure,whichservesasanexecutivesummaryofourliteraturereview’sresults.

FigureB:PotentialObstaclestoSocialAcceptability


February2018


Ourreviewoftheliteraturerevealedsixethicalissues,fallingunderfiveheadings—privacy,socialinclusion, separation of government and business spheres, transparency, reliability and freedom.4Thesecategoriesarebasedonthevarioussubjectsmentionedintheliterature,aswellasexistingclassifications.5

Ourstudyalsohighlightedsevenconcerns,fallingundertwomainheadings:thosepertainingtochangesinmunicipalgovernanceandthosetotransformationofthecityitself.Theseconcernsdonotqualifyasethicalissues,sincetheyarenotclearlyassociatedwithbasicvaluesorprinciples.Yettheyareidentifiedconcernsthatcouldaffectaproject’ssocialacceptability.

Wehavefocusedonfactorsofchange inmodesofgovernance inherentto IoT.Thesefactors,which have major implications for the city administration and residents, are discussed inSection9.Forthesereasons, theyarekeyelementstobeconsidered inanydiscussiononthesocialacceptabilityoftheIoTproject.

EthicalIssues

Sections4to8covereachofthesixidentifiedethicalissues,groupedbythekindsofthreatstheIoTprojectposestothesefundamentalvaluesandprinciples.WemorecloselyexaminesituationsandactivitiesspecifictoIoTthatgiverisetosuchissues.ThesethreatsarealsogroupedbyIoTphase.Forexample,threatstoprivacyindatacollection(suchaslackofpublicconsenttosuchcollection)aredifferentfromthosearisingfromthereleaseofopendata(possibledisseminationofconfidentialinformation).

FigureC,below,outlinestheethicalissuesdescribedinthereport,withrespecttothefourmainphasesofIoTsystem:dataanalysis,releaseofopendata,andpublic,digitalservices.Theseboxespresenttheethicalissuescorrespondingwitheachphaseandthespecificthreatsthatcausetheseissues.Thisreportexplainsallofthesefactorsindepth.

Becauseoftheimportanceofpossiblechangesinthegovernancesystemandtheircoverageinthisreport,theyareidentifiedalongsideethicalissuesinthefollowingfigure.

4Transparencyandreliabilityareconsideredtogether,inthesamesectionofthereport,sincetheysharenumerouscharacteristics.5SuchasthosementionedinareportfortheEuropeanParliament’sLIBECommittee(EuropeanParliament,2015),theETICAProject(Stahl,2011)andtheEU’sEthicsSubgroupIoT(vandenHoven,2016).


February2018


FigureC:SummaryofIdentifiedIssuesandRelatedThreats

3

.

.

Planifica(on*du*projet*de*l’IdO

Collecte*et*stockage*des*données*

1.  Orienta*on,des,inves*ssements,au,détriment,d’autres,enjeux,urbains,

2.  Inégalité,des,inves*ssements,du,point,de,vue,géographique

1.  Sécurité,des,systèmes,et,données,2.  Manque,d’assen*ment,citoyen,pour,la,

collecte,3.  Sen*ment,d’être,surveillé

1.  Façonnement,du,projet,par,intérêts,privés,

2.  Verrouillage,technologique,et,mennotage,contractuel,

3.  Développement,d’une,dépendance,à,la,généra*on,des,revenus,de,vente,des,données,

1.  Sécurité,des,systèmes,et,données,

1.  Tyrannie,de,la,transparence,con*nue,2.  Bonifica*on,des,données,disponibles,

pour,une,cybersurveillance,ubiquitaire,

Inclusion,


Vie,privée,


Liberté,

3

.

.

1.  Généra(on*de*portraits*personnels*(via*combinaison,*ré8iden(fica(on,*inférence)*

2.  Suivi*géographique*3.  U(lisa(on*des*données*pour*des*fins*

différentes*de*celles*communiquées

Analyse(des(données Ouverture(des(données(et(services(citoyens(numériques(

1.  Dissémina(on*de*données*personnelles*confiden(elles*

2.  Dissémina(on*de*données*personnelles*non8confiden(elles*

3.  Mise*à*disposi(on*de*données*qui*peuvent*contribuer*à*générer*des*portraits*personnels*par*autrui

1.  Discrimina(on*par*analyse*algorithmique*

1.  Accès*inéquitable*aux*services*et*données*(fracture*numérique*et*personnalisa(on)*

2.  Accès*inéquitable*à*l’exploita(on*des*données*ouvertes*

3.  Offre*limitée*pour*les*popula(ons*défavorisées*

1.  Offre*de*services*bonifiée*mais*non*publique*1.  Opacité*des*systèmes*et*analyses*

1.  Qualité*des*données*ouvertes*

1.  Analyses*prescrip(ves*orientent*les*choix*des*individus*

2.  Analyses*prédic(ves*décident*de*l’accès*des*individus*à*des*opportunités*

3.  Le*profilage*freine*la*capacité*des*individus*à*se*ré8inventer*

Vie*privée*

Inclusion*

Transparence*et*fiabilité*

Indépendance*des*pouvoirs*publics*

Libe

rté*

Vie*privée*

Inclusion*


1.  Décisions*pour*l’op(misa(on*et*non*l’op(mum*social*

2.  Dilu(on*de*la*responsabilité*décisionnelle*

3.  Déterminisme*4.  Réduc(on*du*champ*des*possibles*Tr

ansforma(

ons*

gouvernance*


February2018


Foreachissuecovered,thereportalsodescribessolutionsidentifiedintheliterature(presentedinFigureD).

FigureD:SummaryofSolutionsIdentified

TheseproposedsolutionsarediscussedinsuccessioninSections4to9,withoutdiscussingthethoughtsoropinionsofthereports’authorsastotheirmerit,maturityorcompatibility.However,thesesolutionssharesomepoints:

● Transparency.● Potentialremedies.● Publicparticipation.● Determinationoftheproject’sbasicprinciples.

.

.

1.  Sécurité*et*vie*privée*dès*la*concep5on*2.  Engagement*à*la*non;ré;iden5fica5on*3.  Collecte*agrégée*et*minimisa5on*des*données*4.  Algorithmes*de*privacité*différen5elle*5.  Solu5ons*en*développement*(ex:"s$cky"flow)*6.  Transparence*et*possibilité*de*recours*7.  Principes*de*base*d’u5lisa5on*des*données*8.  Par5cipa5on*des*citoyens*dans*l’u5lisa5on*des*

données*9.  En5té*publique*de*protec5on*10.  Évalua5on*d’impacts*mul5ples*11.  S’inspirer*des*cadres*existants*12.  S’inspirer*des*champs*de*l’éthique*de*la*

recherche*et*biomédical*

1.  LiTéra5e*numérique:*duca5on*et*accès*aux*technologies*de*l’informa5on*

2.  Promouvoir*pouvoir*poli5que*citoyen,*via*le*numérique*

3.  Services*accessibles*et*per5nents*à*tous*4.  La*loyauté*algorithmique*5.  L’éthique*dès*la*concep5on*6.  Discrimina$on"Aware"Data"Mining"7.  Déterminer*des*règles*pour*encadrer*ce*que*

peuvent*faire*et*ne*pas*faire*les*algorithmes*

1.  Assumer*une*posture*de*pleine*imputabilité*face*au*projet*

2.  Promouvoir*la*par5cipa5on*des*citoyens*dans*le*projet*

1.  Établir*vision*claire*des*besoins*et*valeurs*2.  Générer*les*données*et*en*être*propriétaire*3.  Casser*les*monopoles*

Vie*privée* Inclusion*

Transforma5on*de*la*gouvernance*

Indépendance*des*pouvoirs*publics*

1.  Produc5on*par5cipa5ve*pour*qualité*des*données*

2.  Métriques*et*standards*pour*qualité*des*données*3.  Cul5ver*la*confiance*via*la*transparence*


1.  Respecter*le*droit*à*la*vie*privée*2.  Établir*une*réglementa5on*autour*des*types*de*

prédic5ons*algorithmiques*à*autoriser*et*celles*à*bannir*

3.  Poli5ques*pour*droit*de*contester*4.  Poli5ques*pour*droit*de*ne*pas*être*connecté*et*

oubli*numérique*

Liberté*


February2018


Transparencyisidentifiedindifferentworksasapotentialsolutionapplicabletoalldataphases—collection,use,releaseand(possible)sale,aswellastodecisionsmadefromthedatauseandanybreaches—andappearsseveral timesasapotentialsolution.Naturally, itwouldbedifficultorimpossible to discuss all these topics in detail. However, the idea of transparency on keyprinciples,guidelinesandintentions,constantlyrecursinseveraloftheproposedsolutions.

Givingthepublicaccesstoagrievanceredressmechanismwithrespecttoprivacy,algorithmicbiasandcybersurveillancehasbeensuggestedmorethanonce.Whilethismechanismwouldgivepeopletherighttoaskquestionsandfilecomplaints,itwouldalsocreateaclimateinwhichgovernmentofficialsmustmaintainaudittrailsoftheirdecisionsandprocedures.

Publicparticipation in theurban IoTproject is adominant theme. It isdiscussed in Section4(EthicalIssue:Privacy),whentheauthorsrecommendtransparencyindatacollectionanduse,aswellasinsuggestionsbysomefiguresthatthepublicparticipateintheactualuseofthedata,through information and apps placed at their disposal, giving free reign to creativity andinnovation.Similarly,socialinclusionisseenasakeyfactorinethicalissuespertainingtosocialinclusionitself,andtochanginggovernancesystems.ParticipationisconsideredimportantinallIoTphases—planning,datacollectionandstorage,dataanalysis,andopendataandservices.

Finally,anotherconceptintrinsictoseveralofthesolutionsmentionedistheneedtodefinetheproject’s basic values and requirements clearly. Doing so is crucial for several reasons. Thesectiononprivacydescribesthisneedintermsofdefiningprinciplesandvaluesthatcanguidedata collectionanduse tominimizeprivacy infringements.Delineating theseprinciples is alsoimportant during interactions with the privacy sector in planning and implementing IoTtechnology,wherevendorgoalsmaydifferfromthoseofthecityadministration,orthecommongood.

Privacy

Weweightedourconsiderationofthedifferentethicalissuesintermsoftheirprominenceintheliterature,ratherthanequally.TheissueofprivacyaccordinglydominatesthisreportbecauseofitsimportanceandthethreatIoTposestoit,aswellasitspresencewithinthecomplexexistinglegalframework.

Section4.2,onIoT’slegalframework,startswithareviewoftheprinciplesthathaveguideddatacollectionanduseoverthepastfivedecades.Section4.3thendescribesthecrisisinpersonaldataprotectionduetovariousaspectsofEconomy4.0,suchasthe installationofsensors inpublicareas,accesstovastquantitiesofdataandalgorithmanalysis(Rubinstein,2013;CrawfordandSchultz,2014;Narayanan,etal.,2016;Mantelero,2014;TeneandPolonetsky,2013;Gaughan,2016).IntheparticularcaseofurbanIoTdeployment,thefollowingissuesareofparticularnote:


February2018


● Difficultyofgivingnoticeandobtainingconsentforsensor-baseddatacollection.● Difficulty of ensuring6 that personal data is anonymized, in a context of big data and

predictiveanalytics.● Surgeincreationof“personal”data.● Difficultyofpromotingtheproportionalityprincipleorminimaldatacollection inabig

dataenvironment.Inotherwords,thereisaclearregulatorygapwithrespecttoensuringtheprivacyofIoTdata.

SocialInclusion

Ethicalissuespertainingtosocialinclusionalsoplayakeyroleinthereport.Thissectionexploresthedigitaldivideanditsimpactonpeople’sabilitiestousesmartcityservices,aswellastoengagewiththereleaseddataandplayakeyroleinusingit.

The section also considers the discriminatory potential of algorithm analysis, which profilesgroups for various reasons. Algorithms classify and simplify information according to theirprogrammed values. They are designed to accentuate similarities between members of aparticulargroup,aswellasdifferencesbetweenpreconceivedcategories.Thismeansalgorithmsare inevitably “embedded with values” defined by developers’ operating parameters, asconfiguredbyusers(Mittelstadt,2016).Theliteratureisrepletewithexamplesofsuchalgorithmscausingdiscrimination.

Thesectionalsodiscussesrecommendationalgorithms,whichreferconsumerstonewmarketsby suggesting products, people, services and organizations. Doing so makes it easier to find“similar” information, but can result in unequal access to various opportunities for differentindividuals.

Finally,thereportalsolooksatotherissues(separationofthegovernmentandbusinessspheres,transparency/reliability,andfreedom),althoughsomewhatmoresuccinctly,giventheir limitedcoverageintheliterature.

6Anonymizationeliminatesthelinkbetweenthedataandaspecificindividual(RichardsandKing,2004).


February2018


ChangingModesofGovernance

TheliteratureonsocialissuesassociatedwithsmartcitiesandIoTdoesnotsuggestthatIoTwillresultinbasicchangesingovernance.

Whiletheseissuescannotbequalifiedas“ethical,”theyremainimportantintermsoftheirimpactonday-to-dayadministrationofthecityanditsdealingswithresidents.

Our literature review identifies the factors inherent to this transformation. All relate to thetechnological bias of IoT-based governance, where technology plays a prevailing role ingovernanceanddecision-making.This technologicalbiasengendersadepoliticizationof issuesthesmartcity isexpected toaddress. Inotherwords,wepresent this immenseurbanprojectprimarily as a technological, apolitical and common sense initiative, minimizing debate onproposed political solutions and priorities, and thereby reducing the potential for socialopposition(DouayandHenriot,2016).

Allofthemainfactorsinvolvedinthetransformationofmunicipalgovernancearepresentinthedataanalysisphase.Theyare:

● Decisionsaimedatoptimization,ratherthanthesocialoptimumorrootcauses.● Integrationresultinginareductionorlossofdecision-makingresponsibility.● Deterministicworldviews.● Diminishedrangeofanalyticalscenarios.

Conclusion

TheliteraturereviewgroupsethicalissuesandconcernsabouturbanIoTthatcoulddamagetheIoTproject’ssocialacceptability.

Ourdiscussiongivesequalemphasistothevariousissuesandsolutions.Wedonotwishatthisstage of our work to weigh arguments, solutions or ideas. While we believe this process isessential,itwillbeconductedatalaterphase,incloseconjunctionwiththecity,toensuretheimpartialitythatisexpectedinareviewofthiskind.


February2018


Contents

Summary.........................................................................................................................................4

Introduction....................................................................................................................................4

IoTTechnologicalandAnalyticalFlowchart....................................................................................5

SocialAcceptability.........................................................................................................................7

EthicalIssues...................................................................................................................................8

Privacy...........................................................................................................................................11

SocialInclusion..............................................................................................................................12

ChangingModesofGovernance...................................................................................................13

Conclusion.....................................................................................................................................13

1 Introduction..........................................................................................................................20

1.1 TheProject....................................................................................................................20

1.2 DocumentStructure......................................................................................................21

2 PREFACE:UrbanArchitecture’sHumanImpact....................................................................22

2.1 EvolutionofUrbanInfrastructure:ClassicSocialScience.............................................22

2.2 UsingTechnologicalInnovationtoResolveUrbanChallenges......................................23

2.2.1MoreThanHalftheWorld’sPopulationLivesinCities................................................23

2.2.2CityLifeTransformedbyTechnologicalInnovation.....................................................24

2.2.3OriginandDisseminationoftheSmartCityConcept...................................................24

2.3 OscillatingBetweenUtopiaandDystopia.....................................................................25

3 ScopeandTechniqueoftheLiteratureReview.....................................................................26

3.1 Technique......................................................................................................................26

3.2 TheIoTSystemDiscussedintheLiteratureReview......................................................27

3.2.1 HighlightingVariousComponents.........................................................................29

3.3 Ethical/SocialAcceptabilityIssuesandSolutions:DefinitionsandMethod..................29

3.3.1 EthicalIssues:DefinitionsandMethod.................................................................29

3.3.2 SocialAcceptability:DefinitionsandMethod.......................................................30

3.3.3 ApproachtoSolutions...........................................................................................31


February2018


4 EthicalIssue:Privacy.............................................................................................................32

4.1 WhatisPrivacy?............................................................................................................32

4.2 EvolutionoftheLegalFramework.................................................................................32

4.3 UrbanIoT:LegalFrameworkinCrisis............................................................................34

4.3.1 AdviceandConsentinDisarray.............................................................................34

4.3.2 ThreattoAnonymization.......................................................................................34

4.3.3 ProliferationofPersonalData...............................................................................35

4.3.4 DataMinimization:IncreasinglyLessRelevant.....................................................35

4.4 IoT’sThreatstoPrivacy.................................................................................................36

4.4.1 PotentialThreatsinDataCollectionandStorage..................................................37

4.4.2 PotentialThreatstoInternalDecision-MakinginDataAnalysis...........................38

4.4.3 PotentialThreatsofReleasingDataandProvidingServicestothePublic............38

4.5 PotentialSolutions........................................................................................................39

4.5.1 SecurityandPrivacybyDesign..............................................................................39

4.5.2 CommitmenttoNon-Re-Identification.................................................................39

4.5.3 AggregateDataCollectionandDataMinimization...............................................40

4.5.4 DifferentialPrivacyAlgorithms(DPAs)..................................................................40

4.5.5 PromotingTransparencyandPotentialRemedies................................................41

4.5.6 DefiningBasicPrinciplesofDataCollectionandUse............................................41

4.5.7 PromotingPublicParticipationThroughDataUse................................................42

4.5.8 FormalPublicPrivacyProtectionandImpactAssessmentBody...........................42

4.5.9 DrawingonResearchandBiomedicalEthicalPrinciples.......................................43

4.5.10 DrawingonTraditionalPrivacyFrameworksandtheirEvolution.........................43

5 EthicalIssue:SocialInclusion................................................................................................44

5.1 IoT’sThreatstoSocialInclusion....................................................................................44

5.1.1 PossibleThreatsinIoTInfrastructurePlanningandMaintenance........................45

5.1.2 PotentialThreatstoDataAnalysis.........................................................................45

5.1.3 PossibleThreatsfromDataReleasesandServicesforthePublic.........................46


5.2.1 EducationandAccesstoInformationTechnologies..............................................49


February2018


5.2.2 UnitingPeopleAroundGoals,RatherthanTechnology........................................50

5.2.3 DigitallySupportingthePublic’sPoliticalPower...................................................50

5.2.4 AccessibleandRelevantServicesforAll................................................................51

5.2.5 AlgorithmicAccountability....................................................................................51

5.2.6 EthicsinDesign......................................................................................................51

5.2.7 DiscriminationAwareDataMining........................................................................52

5.2.8 RegulatingWhatAlgorithmsCanandCannotDo..................................................52

6 EthicalIssue:SeparationoftheGovernmentandBusinessSpheres....................................53

6.1 PotentialThreatstoSeparationoftheGovernmentandBusinessSpheres.................53

6.1.1 PotentialThreatstoPlanningandMaintainingIoTInfrastructure........................53

6.1.2 PotentialThreatsinDataAnalysis.........................................................................55


6.2.1 ControllingtheTechnologyandAdministeringPartnershipTerms......................56

6.2.2 Establishing/CommunicatingaClearIdeaofProjectRequirementsandValues..56

6.2.3 GeneratingandOwningData................................................................................57

6.2.4 BreakingupMonopolies........................................................................................57

7 EthicalIssue:TransparencyandReliability............................................................................58


7.1.2 PotentialThreatsofDataAnalysis.........................................................................58

7.1.3 PotentialThreatsofOpenDataandServicesforthePublic.................................59


7.2.1 BuildingTrustThroughTransparency....................................................................60

8 EthicalIssue:Freedom..........................................................................................................62


8.1.2 PotentialThreatsinDataAnalysis.........................................................................64


8.2.1 RespectingPrivacyRights......................................................................................67

8.2.2 BuildingaRegulatoryFrameworkAroundCertainKindsofAlgorithmicPredictionsandDecidingWhichShouldBeBanned................................................................................67

8.2.3 DevelopingPoliciesontheRighttoCommentonandContestData....................67


February2018


8.2.4 DevelopingPoliciesontheRighttobeDisconnected...........................................68

8.2.5 DevelopingPoliciesonaRighttobeForgotten....................................................68

8.3 IsAlgorithmicBiasaNecessaryEvilinEnsuringPublicSafety?.....................................69

9 TransformingGovernanceModes.........................................................................................70

9.1 InternetofThings:FromDriveroftoFactorinChange................................................70

9.1.1 FactorsofTransformationinMunicipalGovernance............................................70


9.2.1 MunicipalAssumptionofFullAccountabilityforProject......................................72

9.2.2 PromotingPublicParticipationintheProject.......................................................72

9.3 DISCUSSION:TheProperRolesforAutomationandAlgorithmsinSociety..................73

10 SocialAcceptabilityandIoT...............................................................................................74

10.1 DefinitionofSocialAcceptability...................................................................................74

10.2 SocialAcceptabilityoftheUrbanIoTProject................................................................75

10.2.1 SocialAcceptabilityoftheSmartCityandSocialFactorsintheSmartCity..........75

10.2.2 Use-PhaseStudies.................................................................................................76

10.2.3 FindingsonSocialAcceptabilityandtheUrbanIoTproject..................................78

10.2.4 IdentifiedConcerns...............................................................................................80

11 Conclusion:TheSmartCity’sPromise...............................................................................82

12 Bibliography......................................................................................................................84

AppendixA....................................................................................................................................94

AppendixB..................................................................................................................................100

AppendixC..................................................................................................................................104

AppendixD..................................................................................................................................108

AppendixE..................................................................................................................................111

AppendixF...................................................................................................................................113

AppendixG..................................................................................................................................115


February2018


Tables

Table1:IdentifiedConcerns.........................................................................................................80

Table2:PotentialSolutionsIdentifiedbutNotExplored..............................................................99

Table3:BasicFIPPs.....................................................................................................................100

Table4:OECDGuidelines............................................................................................................101

Table5:Principlesofthe1990EuropeanDirective....................................................................102

Table6:ExtractofanArticlebyBrentDanielMittelstadt,etal.(2016).....................................111

Figures

FigureA:IoTTechnologicalandAnalyticalFlowchart.....................................................................6

FigureB:PotentialObstaclestoSocialAcceptability......................................................................7

FigureC:SummaryofIdentifiedIssuesandRelatedThreats..........................................................9

FigureD:SummaryofSolutionsIdentified....................................................................................10

Figure1:IoT’sTechnologicalandAnalyticalFramework..............................................................27

Figure2:SimplifiedSysteminQuestion........................................................................................28

Figure3:PrivacyConceptsIdentifiedinVariousLegalFrameworks.............................................33

Figure4:ReferenceModelsofSolove(2007)andZiegeldorf,etal.(2014)..................................36

Figure5:PotentialImpedimentstoSocialAcceptability...............................................................79

Figure6:ContextualPrivacyFactors.............................................................................................95

Figure7:DanielSolove’sClassificationScheme..........................................................................105

Figure8:ModelofZiegeldorf,etal.............................................................................................107


February2018


AbbreviationsandAcronyms

SA Socialacceptability

BAPE Bureaud’audiencespubliquespourtheenvironment

BD Bigdata

CESE Conseiléconomique,socialetenvironmental(France)

CEST Commissionàl’éthiqueensciencesettechnologiesduQuébec

CIRAIG Centreinternationaldereferencesurlecycledeviedesproducts,procédésetservices

DADM DiscriminationAwareDataMining

DPA DifferentialPrivacyAlgorithm

IDF IsraelDefenseForces

FTC FederalTradeCommission

AI Artificialintelligence

ICF IntelligentCommunityForum

IoT InternetofThings

ICT Informationandcommunicationtechnologies

PSTRE Problemsolvingintechnology-richenvironments

RFID Radiofrequencyidentification

SPVM ServicedePolicedelaVilledeMontréal

STM SociétédesTransportsdeMontréal

IT Informationtechnologies

VGI Voluntarygeographicalinformation


February2018


1 Introduction

1.1 TheProject

Since2014,Montréalhasbeenimplementingastrategicinitiativetobecomeaninternationallyrenownedleaderamongsmartanddigitalcities.Montréalplanstodeveloponitsown—andinpartnershipwithresidents—technologicalsolutionstothemetropolis’skeychallenges,whichitwilldeploytransparently,withadvancedtechnologiesandonahumanscale.Thecity’sfourfoldstrategy is to collect/release data to achieve better value for money and promote publicparticipation and innovation, communicate information and foster public connectivity,coordinate digital/smart services for the public, and work with different parties to createnetworksforandacceleratorsofinnovation(VilledeMontréal,2017).

Installationoftechnologicalinfrastructure,andinparticular,thedevelopmentofanInternetofThings in the city, is central to this strategy, which will culminate in the construction oftechnological and analytical systems for data collection, storage and analysis. IoT alsomeansimplanting amultitudeof sensors throughout the city to collect a vast variety of dataon themetropolis’sassetsandactivities.Akey featureof thedigital strategy is thecollectionofdatafrom sensors andother sources for internal useby the city and its partners, accompaniedbyexternalusethroughbigdatareleases.Thecity initiallyplanstofocustheuseofsuchdataonMontréal’s urban priorities, like intelligent trafficmanagement, the environment, urban assetmanagement(furniture,vehicles)andpublicsafety.

While the use of collected data presents opportunities for innovation and for improvingMontreal’squalityoflife,itraisesethicalissuesandrisksthatcouldengendersocialopposition.Inparticular,themereuseofIoT’stechnicalandanalyticframework,alongwiththemanychangesthat will result from new forms of interaction between residents and their municipaladministration,constituteissuesthatmustbeidentifiedandresolved.

ThisreportseekstosupportthecityinitsconsiderationofthistopicbyoutliningpotentialissuesofethicsandsocialacceptabilityinvolvedinusingIoTinthecityandfindingsolutions. Inbroaderterms,itwasproducedunderBatch5oftheInternetofThingsStandardsFormulationProjectandwillserveasastepping-stonetosubsequentprojectphases:


February2018


● Definingaconceptualframeworkfordevelopingaprogramtoanalyze,manageandaddressissuesofethicsandsocialacceptabilityassociatedwiththeIoTproject.

● Establishingthebasisofaproposal forestablishingacityhalladvisorycommitteetodealwiththeethical,legalandsocialramificationsofurbanIoTdeployment.

● Identifyingtopicsandissuesinthisfieldtobeconsideredingreaterdepth.

1.2 DocumentStructure

Thisreporthas12sections.

Section1istheintroduction.

Section2isaprefaceonquestionsconcerningMontréal’sIoTprojectinthecontextofongoingdebateson the greatest revolution in urban architecture inmodern history and its impact on people. TheprefacedescribestheIoTprojectasthetangibleexpressionofarebootedurbanstructureorganizedaroundtheconvergencearisingoutofEconomy4.0.

Section3describesthemethodologyappliedtoourliteraturereviewanddelineatesthescopeoftheIoTsystemconsideredbytheproject.

Sections4to9eachcoverurbanIoTissuesnamedintheliterature,aswellasthechangeingovernancemechanisms. Ethical issues pertain to privacy, social inclusion, separation of the government andbusinessspheres,transparencyandreliability/freedom.EachissueisexaminedintermsofthreatstheIoTprojectposestoourfundamentalsocialvaluesandprinciples.Wealsooutlinepotentialsolutionsidentifiedintheliteraturefordealingwiththeseissues.

Section10defines social acceptabilityanddescribeshow the literaturehas treated theconnectedcity’sissuesofsocialacceptabilitythroughthepresent,aswellaspotentialsolutions.

Section11istheconclusion.

VilledeMontréal Rapportfinal

February2018


2 PREFACE:UrbanArchitecture’sHumanImpact

2.1 EvolutionofUrbanInfrastructure:ClassicSocialScience

Urbanevolutionhas longbeenaclassicsocial sciencetopic.Thecity isan ideal laboratory forresearchersstudyingtheemergenceofnewsocialdynamics,andforthose interested insocialinertia(Fijalkow,2007).

Even inhis day,Georg Simmel (1858-1918) focusedonhow the citydweller’s “mental life” isshapedbyurbandynamics. Inhisnotable1903essay,TheMetropolisandMentalLife,Simmelsought to explain the dramatic changes in personal behaviour due to the emergence of vastmetropolises. At the turn of the 19th century, Simmel lived in Berlin, which was undergoingrelentless,fast-pacedchangeunderthepressureofstrongdemographicgrowth.

Simmel’sverticalhabitat—themetropolis—istheforcedrivingreconfigurationofthesocialrulesofasmalltown(whereeveryoneknowseveryoneelse),andthusalterspersonalbonds.SimmelevaluatedthenewdynamicsofsocialinteractionthroughTheStranger’sperspective.7Basedonthiswork,heultimatelydefinedaphilosophyofurbansociology.

Urbanarchitectureisnot,accordingly,merelythereflectionofasocialsituation,butanengineof social change. We plan to develop this important hypothesis in our discussion onimplementingurbanIoT.

Nathaniel RobertWalker8 recently published an excellent article criticizing his colleagues forunderestimatingindustry’sroleinchangingurbanlifestyles(Walter2016),and,inparticular,theimportanceofcarmakersindefininglanduseplanning.

TheexampleconsideredbyWalkerisofparticularinteresttous,becausehisstudyofmarketingcampaignsfornewurbandesignsinthemid-20thcenturyrevealedthatGeneralMotorswasakeyplayer in this effort.Walker’s research demonstrates that one ofGM’s goalswas tomake allAmericansdependentoncarsbyrestructuringurbaninfrastructure(housing,roads,publictransitandservices),asweseeininEdwardBunker’sNoBeastSoFierce(2016),heraldingtheemergenceofLosAngeles’carcultureofthe1970s.

7InSimmel’sview,citylifeproducestwomutuallysupportingdynamics:(1)individualismand(2)moderntimes.Itdescribestheurbanfabricasaformofsocialconsciousnessandthecityasawayoflife.TheworkofBelgiansociologistJeanRemyupdatedSimmel’sapproachtocontemporarycitiesbyunderscoringtheimportanceofproximityanddistanceinourdealingswithothers(Germain,1997).8AssistantProfessorofArchitecturalHistory.


February2018


WhybeginthisintroductionwiththesereferencestoGeorgSimmelandGeneralMotors?

Both, we believe, provide excellent contexts for studying the connected urban infrastructuretypicalofIoT.AsGeorgSimmelandthenMaxWeber(1864-1920)inhisbookTheCity,9suggested,urban infrastructure implies specific behaviour patterns and it is therefore appropriate toconsiderthesocialimpact(socialandethicalacceptability)ofdigitalurbaninfrastructure.

Furthermore,asWalkermentioned,wemustidentifytheindustrialpurviewofthesmartcity.Isthe smart city simply an industrial project? Simmel wrote that the vertical habitat is thecontemporarycocoon.Inlightofthisclearinsight,weshallreviseurbanconnectivityinitsguiseofamoderncocoon,presidingoverdeploymentoftheconvergencepropellingEconomy4.0.

AFourthIndustrialRevolutionisbuildingontheThird,thedigitalrevolutionthathasbeenoccurringsincethemiddleofthelastcentury.Itischaracterizedbyafusionoftechnologiesthat isblurringthe linesbetweenthephysical,digital,andbiologicalspheres. (Schwab,2016:1).

2.2 UsingTechnologicalInnovationtoResolveUrbanChallenges

2.2.1MoreThanHalftheWorld’sPopulationLivesinCities

Theworldbecameapredominantlyurbansociety in2007when,accordingtoestimates,citiesbecameresponsibleforthreequartersofalleconomicactivity(Brender,2012).Asweknow,citiesonlyoccupy2%oftheglobe’ssurfaceandnowhouse50%oftheplanet’spopulation—afigurethat,accordingtoforecasts,10willleapto60%in2030and70%in2050.Citiesconsume75%ofallenergy produced and are the source of 80% of all CO2 emissions. All sectors associatedwithurbanization (transportation, building construction/maintenance, housing,wastemanagementandenergy)sharetrendsresponsibleforsustainabilityproblems.Manysociologistsbelievesuchproblemsarebreeding grounds for economicdisparities and social exclusion (WilliamWilson,1987;Desmond,2012;Goffman,2009).Furthermore,theincreaseddemographicweightofmajorurban centres gives them greater political and economic clout (Doran, 2014). Cities areincreasinglybehavinglikeautonomousinternationalstakeholders(Olive,2015;LeGales,1995),seekingtobuildontheirreputationsasinnovativehubsamongexperts,themediaandthepublic.

9 As noted by Damien Augias, in “Max Weber et Georg Simmel nous parlent des villes,” Le Monde,28.04.2014.10The2016TheUnitedNationsConferenceonHousingandSustainableUrbanDevelopment(HabitatIII),inQuito,adoptedaNewUrbanAgendatomakecitiesmoreinclusive,secure,resilientandsustainable.Thesmartcitywasthesubjectofaworkingpaper(May31,2005)definingtheroleandstatusofdigitalurbaninfrastructureneededtoaddresssocietalchallenges.http://habitat3.org/wp-content/uploads/21-Habitat-III-Issue-Paper-21_Villes-intelligentes.pdf


February2018


2.2.2CityLifeTransformedbyTechnologicalInnovation

Therearenowover7billioncellphoneaccounts intheworld,comparedtojust738millionin2000.Aroundtheworld,3.2billionpeopleareInternetusers,with2billionofthemindevelopingnations.Mobilebroadbandwasavailableacross47%ofglobe in2015,12 timesmore than in2007.In2015,69%oftheplanet’spopulationhad4Gaccess,upfrom45%in2011for3Gmobiletechnology.Mostaspectsofthenewurbanagendadrawontherolesandabilitiesofinformationandcommunicationtechnologiestomeetgoalsandovercomehurdles(pleaserefertotheHabitatIII Policy Papers), offering the international community new opportunities for and innovativemeansofmakingcitiessecure,resilientandsustainablespacesforeveryone.

2.2.3OriginandDisseminationoftheSmartCityConcept

The firsturbandigital infrastructuredeploymentdates fromtheearly2000s.Launchedby theSouthKoreangovernmentin2003,thesmartcityofSongdowascompletedin2015.11

However,formerUSPresidentBillClintonhasbeencreditedwithpromotingthetermSmartCity.Clinton apparently recognized the convergence of two millennial revolutions: (1) massiveurbanizationand(ii)proliferationofinformationtechnologies.Clintonbelievedthattechnologicalinnovationwouldhelpregulatecityliving:

Theideaseemedtobetheoutgrowthofachallengeproposedin2005bytheformerUSpresident to John Chambers, president of Cisco, manufacturer of digital networkequipment:whynotusetheseamazingresourcestomakecitiesmoresustainable?...Atthe2005SecondWorldSummitofCitiesandLocalAuthoritiesontheInformationSocietyin Bilbao, participants ‘defined a common strategy’ for giving information andcommunicationstechnologiesaccesstotheirterritories.Thiswasthefirsttimethatsuchameeting,organizedbytheUNandtraditionallyreservedforstates,wasopentolocalofficials,privatebusinessesandcivil society . . .Ciscowill study the topic (ona$25Mbudget) andmarket the results in 2010. In 2008, IBM got onboard this first wave ofinvestment (smart cities will be ICT’s biggest customers over coming years), with itsSmarterCitiesinitiative”(Pisani,2015).

11http://songdoibd.com/


February2018


2.3 OscillatingBetweenUtopiaandDystopia12

Urban infrastructure isalteredby installingmultitudesof sensors. Integrationof thesehi-techsystemsintotheurbanhabitattransformsitintoadigitalinfrastructure(Rolland-Villemot,2015).Thedatacollectedfromdigitalsensorsisanewcommodity,withthepotentialtooptimizethemobilityand securityofurban infrastructure,minimize contamination,andcreatenew issues,whileamplifyingoldchallenges.

Based on classical sociological theory (Simmel, Weber and Marx), it is highly likely thatEconomy4.0willinexorablychangecontemporaryurbanlifeinposingnewissuesofethicsandsocial acceptability.Our task is to conduct a sociotechnical study—but the readerwill quicklyrealizethatourfocushasexpandedtoembraceurbanpolicy,aswellasdeliberativegovernance,publicparticipation,protectionofprivacyandthe“informationcommons.”

Readersofthisreportwillalsonotetheemphasiswehaveplacedonadoptingnewdeliberativemechanisms for this groundbreaking IoT technology. A vast number of ideas and potentialsolutionshavebeenproposedtohelpdecision-makersdealwithInnumerablewithoften-thornyproblems.Whilesomeofthesesolutionswillturnouttobe“right,”otherapproachesmightultimatelyprovemore productive. Only time will tell. Faced with the highly interconnected challenges of theAnthropoceneEra,and,beingwrongmoreoftenthanwethink,Mistakes,withacapitalM,arenotonlysuretooccurbutwillbewelcomeifandonlyifmulti-partyassessmentanddiscussionmechanisms are present and functional. Otherwise, the IoT utopia could quickly turn into anOrwelliannightmare.

12 Dystopic literature depicts an imaginary society built around human fearshttps://mondedulivre.hypotheses.org/337


February2018


3 ScopeandTechniqueoftheLiteratureReview

3.1 Technique

Webasedthemethodologyofourliteraturereviewontwocoreissues:

1. WhatethicalissuesareassociatedwithIoTinasmartcity?2. WhatissuesofsocialacceptabilitypertaintoIoTinasmartcity?

WeshouldstartbydefiningwhatwemeanbytheInternetofThingsinMontréal.Wedidsobyconsultingtheliterature,aswellasconductinginterviewswithMontréalcityrepresentatives.Wealso explored our study’s two fundamental concepts—ethical issues and social acceptabilityissues—todeterminetheirscopewithinthisreport.Finally,weconductedresearchfocusingonthesequestions.

Ourliteraturesearchonissuesofethicsandsocialacceptabilityfocusedonscientifictexts(booksandjournals),butincludedsuchgreyliteratureasreportsfromgovernments(primarilyfromtheUS and Europe) and international organizations, as well as blogs by recognized researchers,journalistsandcommentators.Theliteraturecoveredarangeofdisciplines,includingthesocialsciences,computerscience,geography,lawandethics,aswellassuchsubfieldsasurbanstudies,criticaldatastudies,newtechnologyethics,cyberethics,ITethicsandIoTethics.

Weappliedtwoapproaches,oftenhand-in-hand,toourreviewoftheliterature.Thefirstinvolvedkeywordsearches(smartcity,InternetofThings,ethicalissuesandsocialacceptability)relatingtothestudy,indifferentpermutations.13Thenweincludedkeywordsfoundinourreadingsthatpertainmorespecificallytoissuesofethicsandsocialacceptability.Wealsoaddedsuchkeywordsasbigdataandalgorithmanalysis,whichrelatetoIoT’sfundamentaltechnologicalandanalyticalconcepts.

One challenge of the literature reviewwas thatmany relevant sources often did not pertaindirectly to the smart city. Various texts on the ethical issues of IoT considered big data andalgorithmanalysis,butnotintermsofthesmartcity.Inparticular,anumberofsourcesdiscussedIoT in terms of current personal uses (smartphones and FitBit watches), rather than sensorsinstalled by city government. The team producing this report applied common sense inextrapolatingsomeofthisinformationtothesmartcitycontext.

13WesearchedthescientificliteraturewithVirtuose,ScopusandGoogleScholar.GoogleScholarwasusedforgrey literature.Oursecondresearchstrategy involved identifyingreferences inthebibliographiesoftextsconsulted.


February2018


3.2 TheIoTSystemDiscussedintheLiteratureReview

TheliteraturereviewcoveredtopicsspecifictoIoT’sdeploymentinMontréal.Thetechnologicalandanalyticalframeworkoftheplannedsystemappearsbelow,inFigure1.

Figure1:IoT’sTechnologicalandAnalyticalFramework

SteponeisplanningtheurbanIoTinstallation/implementationproject.Next,sensors,aswellashardwareandsoftwareforoperatingthem,willbedeployed.Datawill thenbecollectedfromthese sensors, as well as from external databases, such as social media, and transferred tostorage, and eventually archival, sites. Data will be processed (aggregation, possibleanonymization, etc.) at different sites.While these intermediate steps do not appear on theflowchart, they are important. Data analysis (including predictive and prescriptive) is thenperformed at different points. These analyses primarily serve two key audiences—municipaldecisionmakers and their partners (such as the STM), and the public, such as residents andbusinesses,byreleasingopendataanddevelopingappsforthepublic,designedtoimprovethequalityofurbanlife.

.

.

Installa'on)

Transport)

Stockage)temporaire)

Stockage)ins'tu'onnel)

Analyse)temps7réel)

Analyse)de)flux)

Archivage))

Analyses)interac've)&)en)lot)

Ouverture)des)

données)Analyse)(données)

périodiques)ou)interac'ves))

Visualisa'on)des)données)

et/ou)Résultats)pour)prises)de)décisions)

Analyse)et)développement)d’applica'ons)

par)les)citoyens/entreprises))

Applica'ons)offertes)aux)citoyens)

Légende:()))))))))))))))) )Analyse)des)données)) )Installa'on)des)technologies)et)flux)physique)des)données))

Planifica'on)du)projet)d’IdO)

Destruc'on)données)

Données)captées)

Données)externes)


February2018


Thissystemevolvesinthepresenceofbigdataandnotinavacuum.Newlycollecteddatawill—followingitsreleaseorthedevelopmentofappsforthepublic—becombinedwithexistingbases,frominsideandoutsidethecity,acentralfeatureofthecity’sdigitalstrategy.

Forthesakeofsimplicity,thissystemcanbebrokendownintofourmainphaseshighlightingtheethicalissuesidentifiedinourliteraturereview:

● Infrastructureplanningandmaintenance.● Datacollectionandstorage(includingprocessing).● Internal/externaldataanalysis.● Releasingopendataanddevelopingappsforthepublic.

EachofthesefourstepscorrespondswithoneormoreofthemultiplephasesinvolvedincreatingIoT’stechnicalandanalyticalframework,asshownbelowinFigure2.Dataprocessing(butnotanalysis),forexamplefalls,forthemoment,under“datacollectionandstorage.”

Figure2:SimplifiedSysteminQuestion

.

Version mise à jour

.

Légende'des'grandes'étapes:'!!!!!!!!!!!!!!!! !Planifica)on!du!projet!de!l’IdO!! !Collecte!et!stockage!des!données!

!Analyse!des!données!(interne!ou!externe!à!l’administra)on)!!Ouverture!des!données!et!services!citoyens!numériques!

!

Installa)on!

Données!captées!

Transport!

Stockage!temporaire!

Stockage!ins)tu)onnel!

Analyse!tempsFréel!

Analyse!de!flux!

Analyses!interac)ve!&!en!lot!

Ouverture!des!

données!Analyse!(données!

périodiques!ou!interac)ves)!

Visualisa)on!des!données!

et/ou!Résultats!pour!prises!de!décisions!

Analyse!et!développement!d’applica)ons!

par!les!citoyens/entreprises!!

Applica)ons!offertes!aux!citoyens!

Planifica)on!du!projet!d’IdO!

Archivage!!

Destruc)on!des!données!!

Données!externes!


February2018


3.2.1 HighlightingVariousComponents

Thissystem’stechnicalandanalyticalelementsinclude:

● Technologiesthatgenerateandcollectdata,whiletransportingittomunicipalservers(cams,sensors,RFIDtags,Wi-Fi,etc.).

● Datafrommunicipalsensorsandexternalsources,suchassocialmediaandothersiteswherepeoplecommunicatedirectly(ornot)withthecity.

● Internal/cloudstorageandarchivalsites.● Analyticaltoolsthatconvertdatatoinformation,includingallpredictiveand

prescriptivedataanalysistechniquesinvolvingstatistical,algorithmicandmachine-learningmodels.

● Organizationalstructuresthatsupportcollaborationandinnovationandimprovemunicipalserviceswithnewsourcesofinformation,suchasopendata,permittingconsultationofdataorentiredatabases.14

ItshouldbenotedthatIoTisnotevolvinginisolation,butdrawsonamalgamationsofdata,drawnfromavarietyofsources,tosupportanalysesbymunicipalgovernment,aswellasresidentsandbusinesswithaccesstosuchopendata.AccordingtoMohnanty,etal.(2016),asmartcityisbasedjustasmuchontheuseofsensorsasonbigdata.

3.3 Ethical/SocialAcceptabilityIssuesandSolutions:DefinitionsandMethod

3.3.1 EthicalIssues:DefinitionsandMethod

Asweknow,ethicsisastatementofthecorevaluesandprinciplesthatshouldguideanddirectourinteractionswithothers.Thesevaluesandprinciplesgivemeaningtoourlivesandenableustodistinguish,inagivencontext,betweengood/evil,right/wrong,andappropriate/inappropriate.Inotherwords,anethicalissueariseswheneveramoralprincipleoravaluecomesintoplayinasituation(Commissiondel’éthiqueenscienceettechnologieduQuébec,2017).

Thereareseveralethicaltraditions.MostoftheliteraturewestudiedamplycoversissuesarisingfromIoT,alongwithitsintrinsicandrelatedtechnologies,undertheverybroadheadingofappliedethics, a subfield focusing on tangible situations (Commission de l’éthique en science ettechnologie du Québec, 2017), to which numerous approaches, as well asrecommended/discussedframeworksapply.

14 Adapted fromNational League of Cities, which identified three core components of the smart city.http://www.nlc.org/article/new-report-on-smart-cities-released-by-national-league-of-cities


February2018


In thisproject,wehaveadoptedapluralistapproachaimedat identifying issues raised in theliterature,whateverthedifferentauthors’perspectives.Thisapproach,whichisusedelsewhereaswell(asinEurope’sETICAProject15),shedslightonavarietyofviewpointsandinterpretations,whilehighlightingco-existingoutlooksinthefield(Stahl,2011).

Aspartofthisapproach,welistedthevariousissuesthatwereraised.Next,wegroupedtheissuesin logical categories corresponding to the fundamental values that might be eroded orunderminedbydeploymentofurbanIoT.Thecategoriesareprivacy,socialinclusion,freedom,transparency/reliability, and separation of the government and business spheres. Thesegroupingsnaturallydrewonthoseappearingintheliterature,andinparticular,thosementionedinareportfortheEU’sLiBECommittee(EuropeanParliament,2015),theETICAProject(Stahl,2011)andthewritingsofJeroenvandenHoven(2016).

Weweightedourconsiderationofthedifferentethicalissuesintermsoftheirprominenceintheliterature,ratherthanequally.Theissueofprivacyaccordinglydominatesthisreportbecauseofits importance (van den Hoven, 2016) and the fact that these issues are interwoven withnumerouslegalissuesandframeworks.

3.3.2 SocialAcceptability:DefinitionsandMethod

Socialacceptabilityisacontroversialterm,withmuchdebateoveritsdefinition(Gendron,2014;Battelier,2015).Theexpressiongenerallyreferstotheideathatagrouporthepartiesconcerned(suchas thepublic),consent toaproject thathasbeenofferedto them.AppendixGgivesanoverviewof different definitions for the expression. For this project,we have applied that ofCorinneGendron(2014),whichseemstorepresentacertainconsensusandincludesanumberofconceptsbasedonotherdefinitions.Gendron(2014)definessocialacceptabilityas“thepublicendorsementofaplanordecisionbasedonthecollectivejudgementthattheprojectordecisionisbetterthantheknownalternatives,includingthestatusquo.”Inlinewiththisdefinition,ourreviewoftheliteraturefocusedondocumentsdiscussingpublicconsentforurbanIoTprojectsandtheirinherentorIoT-relatedcoretechnologies.AsmentionedinSection10,thefewstudiesonthistopicdonotatthispointservetoidentify“issuesofsocialacceptability”forthesmartcity.Wecan,however,listethicalissuesandconcernsthatmightgiverisetopublicresistancetowardanurbanIoTprojectandthusconstituteanimpedimenttotheproject’ssocialacceptability.

15Projectonethicalissuesofemergingtechnologies,fundedbytheEuropeanUnion.


February2018


3.3.3 ApproachtoSolutions

Thepotentialsolutionsidentifiedinthisreportare“possibilities”ratherthanactual“solutions”anddescribestrategiesfordealingwithidentifiedissues.Theymayrepresentsolutionsalreadyinapplication, those under development and those merely at the idea stage. They reflect ourfindingsintheliterature,nottheauthors’opinions.Theywereselectediftheymetatleastoneofthefollowingcriteria,althoughmanymetboth:

1. Multiplewritershaveidentifiedthepotentialsolution.2. Itissufficientlyformulatedtopermititsexplanation.

The potential solutions cited in the report do not include all those listed in the literature.Additionaltimeandresourceswouldberequiredtoclarifytheirstatusproperly.

Box1:ImprovedClassificationsofPotentialSolutions

Inthisreport,potentialsolutionstoeachethicalissuearepresentedfollowingourdiscussionofthe issueandaregroupedby“ethicalconcern.”However, itwouldsubsequentlybeuseful togroupthembytheorganizationandwriterproposingeachoption.Itwouldbeespeciallyusefultomapthesesolutionsbygovernment(suchastheEuropeanUnionandUnitedStates).Potentialsolutionsappearingintheliteraturethatarenotwelldeveloped,butaresensible,relevantandgeneric,couldbepresentedinafuturestudy.


February2018


4 EthicalIssue:Privacy

In2014, theUnitedNationsHighCommissioner forHumanRights (UNHCR,2014)publishedawarningon the threats toprivacyofdata collection,preservationand incidentaluse (UNHCR,2014).Manynationalgovernmentsandsupranationalbodieshavesetupspecialcommitteesonthisverytopicoverthepastfewyears,notablywithintheEuropeanUnionandtheUnitedStates.

4.1 WhatisPrivacy?

Therehasbeenmuchdebateabouthowprivacyisdefined.Itcanbeseenaspersonalfreedomfromanyphysicalintrusionintoorinterferencewithaperson’slife(forexample,intermsofhisorherchoices,plansanddecisions),aswellasaperson’sabilitytocontrolaccesstoanduseofhisorherpersonalinformation(Tavani,2004).

Manypeoplemaintainthatprivacyisafunctionofthecontextinwhichinformationisexchanged,rather than the nature of the communication (Nissenbaum, 2004; Barocas and Nissenbaum,2014;16Gaughan,2016).Accordingtothisviewpoint,specific“informationrules”applytoeachexchange,whichiswhypeoplemayormaynotwanttoprotecttheirprivacyinapublicspace,suchasacity.

4.2 EvolutionoftheLegalFramework

Any discussion of privacy must consider its legal framework, which governs personal dataprotection.Thissectionprovidesabriefoverviewofthetopic,whileAppendixBexplores it indetail.

NorthAmericanprivacyprotection legislation is largelybasedon theFair InformationPracticePrinciples(FIPPs)establishedintheUnitedStatesinthe1970s.Theseprinciplesthenevolvedwiththecreationofvariousnationaland internationalpolicies,suchastheOECDGuidelinesontheProtectionofPrivacyandTransborderDataFlows,theFederalTradeCommission’s(FTC’s)PrivacyShieldPrinciplesandtheEuropeanCommissionDirectiveonDataProtection(RichardsandKing,2004;Cate,2006).17

16Nissenbaum(2004,2014)hasendorsedHelenNissenbaum’stheoryofcontextualintegrity(2004,2014),whichAppendix1:Privacyexplainsindetail.17 Recently updated as the General Data Protection Regulation (GDPR), implemented in May 2018(EuropeanParliament,2016).


February2018


ThissetofguidelinesincorporatestheconceptsmentionedinFigure3,above.

.

Figure3:PrivacyConceptsIdentifiedinVariousLegalFrameworks

Manyobserverssaythat implementationstrategiesofgovernmentandbusinessoverthepastfewdecadeshaveunderscoredtheimportanceofadviceandconsent.Protectingtheprivacyofpersonsconcernedbycertaindatameans:1)Informingthesepeopleoftheinformationpracticesof the entity collecting andusing their data and, in particular, specifyingwhat kindof data iscollectedandthereasonsforitsuse.2)Obtainingsuchpersons’consent.18Dataminimizationandanonymizationhavealsobeenproposedasprivacyprotectionstrategies.

18Itshouldbenotedthattheadviceandconsentsystemhasbeenheavilycriticizedformanyyears.Foronething, thenotificationsprovidedareatbestdifficult toreadorunderstandandoftenperceivedasnon-negotiabletermsforaccessingthedesiredservices.Consumersarealsoillequippedtounderstandthem(ViitanenandKingston,2013;Gaughan,2016).


February2018


4.3 UrbanIoT:LegalFrameworkinCrisis

VariousaspectsofEconomy4.0,suchasinstallingofsensorsinpublicareas,bigdataaccessandalgorithmanalysis,havetriggeredacrisisforthepersonaldataprotectionstrategiesdefinedovertheoverthepastfewdecades(Rubinstein,2013;CrawfordandSchultz,2014;Narayanan,etal.2016;Mantelero,2014;TeneandPolonetsky,2013;Gaughan,2016).InthecaseofoneurbanIoTdeployment,thefollowingissuesarose:

● Difficultyapplyingadviceandconsentrulestosensor-baseddatacollection.● Difficulty ensuring anonymization19 of personal data in the presence of big data and

predictiveanalytics.● Generationof“personal”data.● Difficultypromotingprinciplesofproportionateallocationanddataminimizationinbig

dataenvironments.

4.3.1 AdviceandConsentinDisarray

Thedifficultiesoflettingpeopleknowtheirdataisbeingcollectedbysensorsandgivingthemachoiceinthematteriswelldocumented(Schaub,etal.,2015,p.27).Sensorscollectdatafrommovements,emissionsandprocesseswithinthecityonacontinuous,widespreadbasis,withoutany possible timeout to let people “click OK.” Furthermore, predictive data analyses, whichgenerateunpredictableresults,donot lendthemselvestotheconceptofmeaningfulconsent,whichrequiresastatementofpurposeatthestartofdatacollection(TeneandPolonetsky,2013).

4.3.2 ThreattoAnonymization

Development of complex analytic techniques, accompanied by the proliferation of accessible,interoperabledatabases,facilitatesre-identificationofpreviouslyanonymousdata(Ohm,2010;Narayanan,2016;Rubinstein,2013;vandenHoven,2012;EDPS,2014;Gaughan,2016).Advancesincomputerscience,forexample,haveshownthatpeopleinacitycanbeidentifiedwithasfewas four spatial-temporal data points (Montjove, 2013).20 In a 2014 report, the US Council ofScienceandTechnologyconcludedthat ithasbecomeeasier toreversedataanonymization.21Furthermore, the European Data Protection Supervisor noted that data generated by useractivitiesisrarelycompletelyandirreversiblyanonymized(IERC,2015).

19Anonymityisintendedtoeliminatethelinkbetweendataandaspecificperson(RichardsandKing,2004).20Acquisti,GrossandStutzman(2011)havealsodemonstratedthatitisnowpossibletoascertainsocialinsurancenumbersusingaphotoofaperson’sface.AppendixAgivessuchotherexamplesasthescandaloverre-identificationofdatareleasedin2014bytheNewYorkTaxiandLimousineCommission(Tockar,2014;Franceschi-Bicchierai,2015).21This“...couldbeusedasasecuritymeasure,butitisnotonitsowncapableofdealingwithcontemporaryre-identificationmethods”(PCAST,2014,pp.38-39).


February2018


4.3.3 ProliferationofPersonalData

Contemporarydataandtechnologiessupportthecreationofnewpersonaldata.Dataconsideredinnocuousor anonymized in separatedatabasesmaybecome sensitivewhen thosebases aremerged(CrawfordandSchultz,2014;Metcalf,etal.,2016).22Forexample,retailgiantTargetusedits customer database to create analyses identifying women who had recently becomepregnant—evenbeforetheyhadsharedthisnewswithfriendsandfamily(CrawfordandSchultz,2014).

Suchdataamalgamationmaybeviewedasintrusive,whetherornotitpertainstotraditionallypersonalinformation(nameandaddress,etc.).AsTurowaptlystates:“[i]facompanyknows100datapointsaboutmeinthedigitalenvironment,andthataffectshowthatcompanytreatsmeinthedigitalworld,what’sthedifference iftheyknowmynameornot?”(TurowinBarocasandNissenbaum,2014,p.54).

Inanyevent,evenwithoutcollectingandusingindividual-specificdata,inferenceandanalysiscanbe used to assign characteristics to individuals thatmay be related to or identical with dataconsideredpersonal.Suchinferencescanbemadeaboutagroupifasfewas20%ofitsmembersprovideinformationabouttheirpersonalattributes(BarocasandNissenbaum,2015).

4.3.4 DataMinimization:IncreasinglyLessRelevant

Finally,theprincipleofproportionality—orminimizeddatacollectionanduse—thatismentionedinseveralpersonaldataprotectionreferenceframeworks(FIPPs,OECDGuidelines,FTCPrinciples,EUGeneralDataProtectionDirectiveandRegulation)hasbecomeincreasinglydifficulttoenforce,whilethepotentialforinnovationofferedbybigdatarequiresconstantlyincreasingquantitiesofdata(Rubinstein,2013).TheUSFederalTradeCommission,inparticular,believesthatitisdifficulttoenforcedataminimization(ornotificationsandoptouts),withIoT(FTC,2015).

22Somewritersrefertotheconceptofpersonalpublicdata—allinformationnotconfidentialintraditionallegalterms,butthatcanstillbelinkedtoaperson(Tavani,2004;Nissenbaum,1998).


February2018


4.4 IoT’sThreatstoPrivacy

Many of the activities involved in rolling out an urban IoT system could raise ethical issuespertainingtoprivacy.TheworkofZiegeldorf,etal.(2014)andDanielSolove(2007)isparticularlyuseful in identifying potential trouble spots. Ziegeldorf focuses on smart cities and Solove onprivacy,generally.AppendixCtakesacloserlookattheseworks.AsillustratedinFigure4,thetwoauthorsidentifypossiblethreatstoprivacyinvolvedinsuchactivitiesasdatamanipulation(collection,processinganddissemination).23

SolovePrivacyTerminology(2007)

Ziegeldorf,etal.’sModelofThreatstoPrivacy(2014)

Figure4:ReferenceModelsofSolove(2007)andZiegeldorf,etal.(2014)

Based our research and that of the other writers we have consulted, we have presented asummaryofthreatstoprivacy24associatedwiththeseethicalissues,below.ThesethreatspertaintothreeofurbanIoT’sfourcornerstones:

● Datacollectionandstorage.● Dataanalysis.● Datareleasesandservicesforthepublic.

23Theauthorsidentifiedotheractivities.Solove(2007)mentioned“invasion,”whichisnotrelevanttoIoT.Ziegeldorf,etal.(2014),referredtointeractionandpresentation.24These“threats”areactivities,situationsandcontextsthatcouldcompromiseprivacy.


February2018


4.4.1 PotentialThreatsinDataCollectionandStorage

Thisphaseincludesdatacollection,transmission,storageandarchiving.Threethreatstoprivacyhavebeenidentifiedateachofthesesteps:

● Datasystemsecurity.● Personaldatacollectionwithoutconsent.● Thepublic’ssenseofbeingundersurveillance.

DataandSystemSecurity

Theexposureofcomputersystemstocyberandphysicalattackshasbeenwidelyreported(IERC,2015;FTC,2015;vandenHoven,2012).AstudycommissionedbytheEuropeanParliamentlistsdatabreaches,malwareinfections,unauthorizedaccesstopersonaldataandillegalsurveillanceas such vulnerabilities (Euro Parl, 2015). The US Federal Trade Commission has named twoprincipalweaknesses.Thefirstislikelybreachesofthecentralcomputingsystemthatcouldgivehackerspersonalinformationthathasbeenstoredin,ortransmittedbyasensor,toit.Thenthereare in sensor securitygaps that could facilitateattackson thenetwork towhich the sensor isconnected(FTC,2015,p.26).

The omnipresence of sensors in a connected city also amplifies vulnerability. Such sensorsincrease hacking opportunities, particularly as they often lack the capacity to containsophisticatedsecuritysystemsandit ishardtodetectattacksbecausetheyaresosmall(IERC,2015).Writershavealsonotedthatbigdatastoragemagnifiespotentialdamageduetoanydatabreachofsensors,serversorcloudstorage(FTC,2015).

PersonalDataCollectionWithoutConsent

Lackofinformationornoticeaboutthecollectionandplanneduseofsensordatacanheightenpeople’sconcernsthatdataisbeingcollectedwithouttheirknowledge.Manywriterscoverthisissueintheliterature,oftenwithrespecttotheInternetofEverydayThingsandonlineprivacy25(Tavani,1999;BarocasandNissenbaum,2014;Solove,2007;ViitanenandKingston,2013).

FeelingWatched

Thesenseofbeingundersurveillanceduringdatacollectionmaycontributetoprivacyconcerns(Solove,2007).Section9considerssurveillanceanditsimpactonfreedomandpersonalidentityindetail.

25Inhisclassificationscheme,Solove(2007)describesthisissueinboththecollectionanddataprocessingphasesasan“exclusion.”However,webelievethethreatisonlypresentduringcollection.


February2018


4.4.2 PotentialThreatstoInternalDecision-MakinginDataAnalysis

Thefollowingthreatstoprivacyarepresentatthisstep:

● Personalprofilesgeneratedbymergingdata.26● Personalprofilesgeneratedthroughre-identification.● Personalprofilesgeneratedthroughprofiling(inference).● Geographictrackingofpeople.● Datausedforpurposesotherthanthosestatedandintended.

GeneratingPersonalProfilesandGeographicTracking

As discussed in Section 4.3.3, it ismore likely than ever before that personal profileswill begenerated throughcross-matchingdataandprofiling (Section5examines the latter indepth).Ziegeldorfhasalsoreferredtothepotentialrisksarisingfrompersonalgeolocation.

UsingDataforOtherthanStatedPurposes

While it is not easy to apply advice and consent requirements to urban IoT, harmoniousdeploymentofsuchtechnologies impliessomekindofpublicconsenttothecollectionofdataand its intended purposes. Data uses that are other than stated or clandestine have beenmentionedaspotentialinfringementsofpersonalprivacy.

4.4.3 PotentialThreatsofReleasingDataandProvidingServicestothePublic

Fordatatobecomeopen,itmustbereleasedbythoseadministeringit.Suchdatamaythenbeexamined,analyzedandusedbythirdparties.Threethreatsarepresentatthisstage:

● Distributionofconfidentialpersonaldata.● Distributionofdataconsiderednotconfidentialunderthelaw,butwhichcouldaffect

someone’sreputation.● Accesstodatathatcouldcontributetothegenerationofpersonalprofilesthroughdata

combination,re-identificationandprofiling.

It isclearthateachdatarelease increasesthequantityofexistingdataandcontributestothelikelihood thatpersonalprofilesareproduced throughdatacombination, re-identificationandprofilingbythirdparties.

26 Includes Solove’s concepts of aggregation and identification, and Ziegeldorf’s of identification andlinkage.


February2018


4.5 PotentialSolutions

PrivacyconcernspertainingtoIoTandthesmartcityarecomplex.Whiletheliteraturedoesnotofferanymiraclesolutions,certainpotentialoptionshaveemerged,particularlywithrespecttotechnological,politicalandlegislativeremedies.

SeveralIERCteamshavebeencreatedoverthepastfewyearstodeveloptechnicalsolutionstoIoT security and privacy issues, including data confidentiality (through sticky flow policies27),metadata anonymization, user data anonymization, cyber-physical security, hardwareauthenticationandequipmentidentitymanagement.Formoreinformationonthisinitiativeandothersunderdevelopment,seeIERC,2015.

4.5.1 SecurityandPrivacybyDesign

Thisapproach,advocatedbymanywritersandleadorganizations,encouragessystemdesignerstoincorporatesecurityandprivacyfeaturesinthedevelopmentandconstructionofsensorsanddataprocessingsystems(vandenHoven,2012;FTC,2015;RichardsandKing,2004).Previously,such considerations were considered to fall within the exclusive domain of policymakers(Gaughan,2016).28

4.5.2 CommitmenttoNon-Re-Identification

Anonymization is standard practice among European and North American regulatory bodies,althoughitisnotconsideredadequateonitsown.29Theliteraturedescribesvarioussolutionsforovercoming the limits of anonymization, including organizational commitment tonon-re-identification.

27ResearchersfundedbytheEuropeanCommissionarenowseekingtoapplystickyflowpoliciesensuringthataccessandconfidentialityrulesareassociated,throughmetadata,withdataflows.Thisincludestheideathateachdatapointinasystemispairedwithasecuritypolicydefininghowthedatacanbeusedandwhichconditionsmustbemetbeforeitcanbemigratedtoanewdataprocessingunit(IERC,2015).28AppendixAdescribesthemainlinesofthisapproach.29 IntheEuropeanCommissionDirectiveonDataProtection95/46/EC,requiringthatalldatasourcesbeanonymized or scrambled with a quantitative privacy guarantee corresponding to the likelihood ofre-identification(vandenHoven2012).ThesamerulehasbeenappliedontheothersideofthePond,withPCASTproposinganonymizationasastrategytodedeployed,butwhichisinadequatebyitselftoprotectdata(PCAST,2014).


February2018


Ina2015report,theFederalTradeCommissionrecommendedthatbusinessesstoretheirdatainanonymizedform(FTC,2015)andimplementmechanismstoensurelong-termde-identification.Accordingtothisapproach,businessesshould:1)Takereasonablemeasurestode-identifydata.2)Publiclycommitnottore-identifydata.3)Concludebindingcontractswiththirdpartiessharingthedata,includingthelatters’commitmentnottore-identifyit(FTC,2015).

4.5.3 AggregateDataCollectionandDataMinimization

Coarse-graindatagathering is anotherway toovercome the limitsof randomization (vandenHoven,2012).Gaughan(2016)gavetheexampleofdatacollectionsystemsthatcanbeconfiguredtoaggregatedataatthesource,preventinglocaldisaggregateddatafrombeingassociatedwithindividuals (Gaughan, 2016, p. 60). In many cases, aggregate statistics appear to meet theanalyticalneedsofcitygovernments.30

Finally,many frameworks are designed tominimize data collection and use, according to theprincipleofgatheringonlynecessarydataandkeepingitmerelyforthetimestrictlynecessaryforanalysis. This approach reduces thequantityof data that couldbeused to generatepersonalprofilesandminimizevulnerabilityincaseofabreach.

ItshouldhoweverbenotedthataggregateandminimizeddatacollectionarestrategiesthatruncountertothefundamentalgoalsofurbanIoT,aswellastoonegoalofanIoTproject—collectinglargequantitiesofdisaggregateddata,topromotesubsequentreuse.

4.5.4 DifferentialPrivacyAlgorithms(DPAs)

DPAsgenerate“static”(small,quantifiableerrors)inanalyticalresults,makingthem“fuzzy,”incontrastwithresultsbasedonoriginaldata(Narayanan,2016).AsNarayananexplained,DPAs,like all data protection measures, represent a compromise between protecting privacy andmanipulatingdataeasily.TheUSCensusBureauemployedthisstrategytoprotectprivacyinitsOnTheMapprogram,asdidGoogleforitsRAPPOR(RandomizedAggregatablePrivacy-PreservingOrdinalResponse)initiative(Erlingsson,etal.,2014).

30AppendixAincludesmoreinformationontheseapproaches.


February2018


4.5.5 PromotingTransparencyandPotentialRemedies

Severalexpectsrecommendfocusingonhowcollecteddataisusedandthedecisionsresultingfromsuchuse(TeneandPolonetsky,2013),aswellasonthedataitself(EuropeanParliament,2015).Thisisparticularlyimportantwheremoretraditionalformsofconsentdonotapplyand“othertrustandtrustedprocedureshavetobeimagined”(IERC,2015).

DrawingonDanielleCitron’s(2010)31workonautomateddecision-making,TeneandPolonetsky(2013),Rubinstein(2013)andCrawfordandSchultz(2014)haverecommendedthatorganizationsbe required to reveal thedataused, criteriaand inherent logicof theiranalyticalprocessandprocedures for making decisions using the collected data. The authors argued that such arequirement could discourage improper profiling and give people the opportunity to appealdecisionsmadebyalgorithm-basedprocesses.This systemwouldrequireentities thatanalyzedatatoproduceanaudittrail(CrawfordandSchultz,2014,p.33).32

Transparency is also strongly recommended in data sales (Herschel and Miori, 2016).Furthermore, the principle of mandatory notification following a security breach has beenintroducedintheUnitedStates(FTC,2015),andintheupdatedEUDataProtectionRegulation(EuropeanParliament,2015).

Naturally, transparency isalsorecommendedasapossiblesolution inSection7 (Ethical Issue:TransparencyandReliability).

4.5.6 DefiningBasicPrinciplesofDataCollectionandUse

Improving transparency indatacollectionandusemeansestablishingguidelineson the topic.Suchguidelinesshouldbeclearbutflexible,tomakewayforthefutureevolutionoftechnologiesandperceptions(HarvardLawSchool,etal.,2017).VandenHoven(2012)notedtheimportanceofstipulatingwhocanaccesswhichinformation,underwhatcircumstancesandforhowlong,aswellashowtheinformationcanbeusedandcombinedwithinformationfromotherdata(VandenHoven 2012). Some cities have begun setting basic universal privacy rules, particularly intermsofsupplierrelations33(HarvardLawSchool,etal.,2017).Seattle’sinitiativeinthisareaisoneexampleofthatphenomenon.

IdentifyingbasicprinciplesisalsoapotentialsolutionrecommendedinSection6(EthicalIssue:SeparationoftheGovernmentandBusinessSpheres).

31Citron(2010)alsosuggestedseveralsupplementarymeasuresthatcouldbeappliedinthisarea,suchasinvesting in unconscious bias training andmistakes in automated decision-making for employees usingthese decision-making systems (Citron, 2010). More information on Citron’s recommendations can befoundin(2010)AppendixA.32AdditionaldetailsontheProceduralDueProcessapproachrecommendedbyCrawfordandSchultz(2014)appearAppendixA.33Thereportdoesnot,however,statewhattheserulesare.


February2018


Box2:SeattlePrivacyPrinciples(2015)

Followingmonthsofconsultationswithstakeholders,Seattleadoptedsixprivacyprinciples inFebruary2015.Theyare:

1. Wevalueyourprivacy:Privacyimpactassessmentswillbeconductedonallnewdataprograms.2. Wecollectandkeeponlywhatweneed:Thecityonlycollectstheinformationitneedstodeliver

cityservices.3. Howweuseyour information:Whenpossible, thecitymakesavailable informationabout the

waysitusespersonalinformationandcommitstogivingpeopleachoicewheneverpossibleabouthowitusestheirinformation.

4. Weareaccountable:Thecitycomplieswithallfederalandstateprivacylaws.5. How we share your information: The city follows federal and state laws about information

disclosure.Businesspartnersandcontractedvendorsthatreceiveorcollectpersonalinformationfromthecitymustagreetoitsprivacyrequirements.

6. Accuracyisimportant:Thecityworkstocorrectinaccuratepersonalinformation,whenpractical.

Thecityconsequentlyadoptedaprivacycommitmentbasedonthesesixprinciples,spellingoutprivacyanddatamanagementpracticesforallitsdepartments.Thiscommitmentalsorequiresaprivacyimpactassessmentandaprivacythresholdanalysisforallnewdatacollectionprograms(Gaughan,2016).

4.5.7 PromotingPublicParticipationThroughDataUse

TeneandPolonetsky(2013)advocate“sharingthewealth,”basedontheideaofgivingpeopleaccess to their data in a useful, attractive format. Proponents of this approach believe thatencouraging interaction between people and data is oneway of bringing them on board thedebateoverdataandtheirrightsinthisarea.Theyalsoseethisstrategyasawayofpromotingexpansionofthedataecosystem,solutionsandpublicapps,enablingpeopletoanalyzetheirowndata and arrive at useful conclusions. As an example, the authors mention the Obamaadministration’sGreenButtoninitiative,whichgivesconsumersaccesstotheirownenergyusagedatainauser-andcomputer-friendlyformat(TeneandPolonetsky,2013).

4.5.8 FormalPublicPrivacyProtectionandImpactAssessmentBody

A number of writers have supported the creation of a public privacy protection organization(Solove, 2007;Mantelero, 2014), similar to thosenowenforced inquasi-public and consumerprotection organizations. Mantelero (2014) said that such an entity should have access totechnological knowledge and a broad societal perspective to evaluate risks posed by datatreatment/analysis,andbalancetheinterestsofdifferentstakeholdersthatcouldbeaffectedbylarge-scaledatacollection,extractionandanalysisprojects(Mantelero,2014,p.19).


February2018


Mantelero (2014) also recommended conductingmultiple robust impact assessmentsondataprocessing,includingtheopportunityforpeopletooptoutofcertainanalyses.Itshouldbenotedthat privacy impact assessments have existed since the 1990s.34 This new model, however,introducestheideathatorganizationsusingbigdatamustconductanimpactassessmentofdataprotection and surveillance, as well as possible discrimination in the analyses and developappropriatemeasurestominimizesuchdiscrimination.Theseassessmentsshouldbeperformedby thirdpartiesandsupervisedbydataprotectionofficialswhowouldalsobe responsible fordefiningassessmentcriteria(Mantelero,2014,p.26).

4.5.9 DrawingonResearchandBiomedicalEthicalPrinciples

Researchandbiomedicalethicalpracticesmayprovideusefulframeworksforsolutions.Barocasand Nissenbaum, for example, have proposed instilling ethical rules, much like a doctors’HippocraticOath,amongscientificanalystsandresearchers.Theyalsorecommendusingethicsreview boards to evaluate data collection, use and dissemination programs (Barocas andNissenbaum,2014).35

The Council for BigData, Ethics, and Society, consisting of 20 universities prominent in thesesocial,naturalandcomputersciencedisciplines,hasformulatedanethicscodebasedonPLOSComputationalBiology’s“Tensimplerulesforresponsiblebigdataresearch”(Zook,etal.,2017).Thefirstfiveconcernminimizingpotentialdamagecausedbybigdataresearchandtheremainderpertaintotheuseofbestpracticesbyresearchersintheirwork(Zook,etal.,2017).The10rulesappearinAppendixA.

4.5.10 DrawingonTraditionalPrivacyFrameworksandtheirEvolution

Several writers and stakeholders agree that traditional privacy principles are not completelyapplicable to contemporary IoT.However, some authors do agree that these principles couldguidedecisionsinmanywaysandshouldsimplybeupdatedsotheycancontinuetofulfilltheirpurpose.

34TheFederalTradeCommission,forexample,recommendsthatbusinessesconductsecurityandprivacyriskassessments(FTC,2015,p.44).Theorganizationalsorecommendsthatbusinessestesttheirsecuritymeasurespriortodeployment.35However,somewritersseemanylimitationsinthisproposal, includingthefactthatdatascienceasawhole(ratherthanaparticularprofession,suchasmedicine)employsavarietyoftoolsindifferentcontexts(private/publicsector,etc.)(Gaughan,2016,p.53).Suchacommittee’sauthoritycouldbecompromisedbyalackoflegalremedies(suchaslicencesuspension).


February2018


5 EthicalIssue:SocialInclusion

Many municipal actors contend that new technologies should be employed to reduce—notaccentuate—socialinequality(HarvardLawSchool,etal.,2017).YettheinherentnatureofIoTand the context of its deployment indicate that there aremany hurdles to overcome beforeachieving these goals. Furthermore, our literature review underscores multiple ethical issuespertainingtosocialinclusion.

Box3:Discrimination

This section covers the concept of discrimination—attitudes based on stereotypes andprejudices.36AmbroseBierce37definedprejudiceas“avagrantopinionwithoutvisiblemeansofsupport.” In other words, prejudice is a groundless assertion. Stereotypes are rudimentarydescriptions and rigid simplifications used to characterize a thing or person. They are allembracing,prefabricated,socialandusedinavirtuallyautomaticorroutinemanner(Moscovici,2014).

Stereotypes and prejudices are used to describe the world in an oversimplified way, byoverstatingsimilaritiesbetweenmembersofaparticulargroupanddifferencesbetweengroups.PleaseseeAppendixDforfurtherinformationonthissubject.

5.1 IoT’sThreatstoSocialInclusion

Possiblethreats38tosocialinclusionmayoccurinallfourphasesofurbanIoT:

● IoTprojectplanning.● Datacollectionandstorage.● Internal/externaldataanalysis.● Releasingopendataandprovidingservicestothepublic.

36Stereotypesandprejudicearetwoexpressionsofacollectivecharacterizationthatdefiesanalysisandfocusesondescribingothers(Amossy,1989).37TheDevil'sDictionarycontainsninety-eightsatiricaldefinitionswrittenbyAmbroseBiercefrom1881to1906.38These“threats”areactivities,situationsandcontextsthatputprivacyatrisk.


February2018


5.1.1 PossibleThreatsinIoTInfrastructurePlanningandMaintenance

TwothreatstosocialinclusionhavebeenidentifiedattheIoTtechnologicalinfrastructureplanningandmaintenancephase:

● ChannellinginvestmentstowardIoTratherthantootherurgenturbanissues.● Geographicallyunequalinvestment.

Aconcentrationofresources inthesmartcityproject,accompaniedby thewidespreaduseofsmartcityrhetoric,promotingtheconceptasinherentlybeneficial,apoliticalandthusessentiallybeyondcriticism—hasservedtochannelmassivepublicinvestmenttowardITratherthantoothernecessary action areas (Söderström, 2014). For this reason, multiple stakeholders havecondemnedthe failure toaddresssuchhigh-priorityurbanproblemsaspropertyprices, socialharmony/relationshipsbetweendifferentcommunities,andretentionof local services,etc., inrollingoutthesmartcityasaservice(Kaplan,2012).Inthecaseofthesmartcity,specialemphasishas been placed on changes in such areas of governance as energy, transportation andtechnology—tothedetrimentofothers(Felli,2015).

Concentrationsoffunding,infrastructureandpoliciesarealsoshiftinggeographically,withsomeurbandistrictsbenefittingtothedetrimentofothers(Felli,2015).DeBreuxandDiaz(2016)alsonotedthatstudiesonsuchself-proclaimedsmartcitiesasSingapore,RiodeJaneiroandBoston,presentamoremundanesituation,whereurbanintelligenceisconfinedtoafewsmalldistrictsand certain sectors of governance. The same rules apply to nearby suburbs,which are oftenexcludedfromsuchtechnicalandsocialprojects,increasingthesocialgapbetweencommunities(Felli,2015).

Finally,somewriterssaythatamajorprojectofthiskindwouldobfuscatetheslightprogressofotherurbanplanninggoals,aswellascontroversiesoveractualbenefitsofcertainvalue-addedprojects(DouayandHenriot,2016).

5.1.2 PotentialThreatstoDataAnalysis

Onethreattosocialinclusionhasbeenidentifiedatthisstep:

● Discriminatoryprofilingbyalgorithms.Manystudies refer to thepotentiallydiscriminatoryeffectofalgorithms.RomeiandRuggieri(2013)haveestablishedacomprehensiveinventoryofthevastnumberofdiscriminationsurveysinvolvedindataproductionandanalysis.Theirarticleidentifiesanddescribesavarietyofpitfallsthathavenotonlyskewedtheproductionofscientificdata,butreinforcedexistingformsofbias.Furthermore,theymentionedthatalgorithmicsoftwareemploystereotypesandprejudices. Inmanycases,suchappsevenservetoamplifysocial,legalandeconomicdiscrimination(Sweeney2013;BarocasandSelbst2015;Birrer2015;Winter,2015),oftenresultinginunfairtreatmentofindividuals,basedontheirprofiles(Goodman,2016).


February2018


Algorithms classify and simplify information according to their programmed values. They aredesignedtoaccentuatesimilaritiesbetweenmembersofaparticulargroup,aswellasdifferencesbetweenpreconceivedcategories.Thismeansalgorithmsareinevitably“embeddedwithvalues”definedbydeveloperoperatingparametersandasconfiguredbyusers(Mittelstadt,2016) ,aswellasthroughpossiblediscriminatorybiaspresentindata,whichoftenreflectsexistingsocialstereotypes.Poorqualitydatacanalsocreatealgorithmicbias.

Box4:TheExampleofCOMPASSoftware

American judges use the COMPAS algorithm39 in sentencing. To determine its effectiveness,ProPublica(Larson,etal.,2016)comparedthesentencesof10,000peoplearrestedinBrowardCounty,Florida,withthealgorithm’spredictions,whiletheyweredetainedin2013and2014.The journalists counted the number of previous defendants who were rearrested over thesubsequent two years. The large number who were clearly illustrated ethnic discrimination.AfricanAmericanswere twiceas likely tobe incorrectlyperceivedaspotential violent repeatoffenders.Whiteswhore-offendedandhadpreviouslybeenchargedwithviolentcrimeswere63%morelikelytobeincorrectlyassignedalowlikelihoodofviolentre-offensivewithrespecttoablackcriminalhavingthesameprofile.

5.1.3 PossibleThreatsfromDataReleasesandServicesforthePublic

Fourthreatstosocialinclusionhavebeenidentifiedatthisstep:

● Limitedaccesstodataandservices,duetothedigitaldivide(thosenotabletoaccessorunderstandthedata).

● Leadroleforthepublicinusingopendataimpossibleorlimited.● Limitedurbanaccessfordisenfranchisedtargetgroups.● Unequalaccess,dependingondigitaluserprofile.

UnequalAccessDuetotheDigitalDivide

Aspreviouslymentioned,theurbanIoTprojectaimstoimproveservicestothepublic,notablythroughdevelopmentofeffectiveonlineservices (includingapps)andopendata,whichcouldstimulate the creation of new studies and services for and by the public. Under thesecircumstances,maintaining telephone-based and face-to-face services is important for groupswithlessInternetaccess,inviewofcurrentprogressinenhancingonlineservices.

39CorrectionalOffenderManagementProfilingAlternativeSanctions.


February2018


Givingthepublicaccesstoupgradedmunicipaldigitalservicespresupposesthatitcanunderstandthesetechnologiesandtheinformationtheyprovideandusethemtoimprovetheirqualityoflife(Poty,2014).Theseservices,inshort,areonlyaccessiblebyindividualswithreading,40numeracy41andproblem-solvingskillssuitedtotechnology-richenvironments(PSTRE).42This“digitaldivide”constitutesamajorchallengetosocialintegration(RalletandRochelandet,2004;Mossenburg,etal.,2003).Thatissuedovetailswithandcanreinforceexistingtraditional(wealth),demographic,territorialandeducationalsocialdivisions(Peres,2015).

ThedigitaldivideispresentinMontréalandthroughouttherestofQuébec.AccordingtodatafromtheInstitutdeStatistiquedeQuébec,aboutoneinfiveQuebeckershaspoorliteracyandnumeracy skills.43 Furthermore, 51% of the population has poor to very poor Internet skills.Finally, 16.4% of Montréal households had no Internet connection in 2012 (Institut de laStatistiqueduQuébec,2015).

Such difficulties accessing municipal services generate inequality: “Those with limited accesseitherpaywithtimelosttryingtousetheseservicesoringoingtospeakwithsomeonedirectly.A worse scenario is simply not using certain services out of frustration” (Conseil national dunumérique,2013,p.78).Furthermore,reducedaccesstoservicesandinformationavailableto“the others” creates a sense of exclusion, worthlessness and powerlessness amongdisenfranchisedgroups(Plantard,2013).

40Theabilitytounderstand,assess,useandmakecommitmentsintexts,toplayaroleinsociety,achievepersonalgoalsanddeveloppersonalknowledgeandpotential(OECD,2014,p.20,inISQ,2015).41Theabilitytofind,use,interpretandcommunicateinformationandmathematicalconceptstodealwiththemathematicalrequirementsofmanysituationsinadultlife(OECD,2014,p.20,inISQ,2015).42 Using digital technologies, communication tools and networks to acquire and evaluate information,communicatewithothersandperformpracticaltasks(OECD,2014,p.32,inISQ,2015).43“Poor”literacyornumeracymeansan“inferior”skilland“Level1”oftheISQScale.InthecaseofPSTRE,“verylowlevel”means“inferior”and“low”means“Level1”oftheISQScale.PleaseseeAppendixFforfurtherinformation.


February2018


LimitedAccesstoLeadRoleUsingOpenData

Onlythose individualswithsolid literacy,numeracyandPSTRE(ProblemSolving inTechnologyRichEnvironments)skillswill,atleastovertheshortterm,beabletoplayimportantrolesinusingopendataforanalysis,enrichingdemocraticdebateanddevelopingnewapps.Suchrolesnotonlyrequireanunderstandingofandacontroloverbasictechnologies,buttheabilitytoanalyzedataand develop programs. This means revising the concept of digital divide beyond the currentnarrowfocusonuserskills,whileconsideringtheimportanceofbeingabletoanalyzeandusedata.

FewMunicipalServicesforDisadvantagedGroups

Becauseoftheirdifficultyaccessingonlineservices,digitallymarginalizedgroupshavelesschanceofgettingtheirneedstorankhighlyamongnewlycreatedservices(ViitanenandKingston,2013).Such needs pertain to technology and content. Technologically, the vast majority of onlineservicesaredesignedforanaudiencewithfunctionalliteracy,numeracyandPSTREskills—ratherthanthoseatthebottomofthelearningcurve(SantaClaraUniversity,2017).Intermsofcontent,the development of services targeting a digitally literate audience is bound to disregard thespecial needs of disenfranchised groups. This is all the more important in view of overlapsbetween digital disenfranchisement and of other social determinants of marginalization(education,wealth,demographics,etc.).Thatmeanstwoforcesareatworkaccentuatingurbansocioeconomic disparities (Viitanen and Kingston, 2013), with services provided in a formatinconsistentwithpersonalabilitiesandpoorlyequippedtomeetspecialneeds.

UnequalAccessforDifferentUserProfiles

Algorithmsemploythedigitaltrackspeopleleave,knowinglyornot,inmostappstheyuse.Avastarray of algorithmsuse data tracks (words, sounds, images, numbers, descriptors) to channelconsumers toward new markets (recommended products, services and organizations). Thisresultsinbothdiscriminationandisolation,sinceusers,dependingonthetrackstheyleave—thatare a function of their social class, inquisitiveness and education—will have very differentopportunities,whichinturntendtoreinforcetheirowndiscriminatorybiases.

Box5:WilltheSmartCitybeaGhettooftheRich?

Inclusionisthefocusofvarioussocialmovements.OnFebruary8,2017,Ouishareorganizedaroundtableon“SMARTCITIESDON’TLIKETHEPOOR,”whichtookthepositionthatsmartcities“will become a ghetto of the rich.” The Ouishare community, in response, supports publicingenuity and recommends replacing themarketing posture of “Smart” with the concept of“sharedcityliving.”


February2018


Box6:PublicParticipation,byCivicTechInCanada’sEnglish-speakingcities,ledbyToronto,theterm“CivicTech”44istheumbrellatermforallofamunicipality’spublictechnologicalinitiatives.Usingsharedoropen-sourcedata,appsarecreatedforusebythepublic.Forcivictechnologyproponents,theseappsconstituteaseachangeinthepublicparticipationmodelbyemphasizingthe“logicofaction.”TheannouncedintentionofrepresentativesoftheinternationalandurbanCivicTechmovementistogivepower,inadditiontothevote,backtothepeople.CivicTechismakingcity-dwellerstruestakeholdersinthecommongood.


Theliteratureandinterviewsidentifythefollowingpotentialsolutionsforalleviatingethicalissuesofsocialinclusion:

● Emphasizingeducationandaccesstoinformationtechnologies.● Gettingresidentsonboard,byfocusingonmunicipalgoalsratherthantechnology.● Makingdigitalliteracyapolicychallenge.● Developinguniversallyaccessibleandusefulservices.● Developingopen,transparentalgorithms.● Developingalgorithmsthatcomplywithcertainethicalprinciples.● More effectively including external partners (media and advocacy groups) that are

engagedinthedisseminationofinformation.

5.2.1 EducationandAccesstoInformationTechnologies

Anumber of observers advocate support for digital education (Peres, 2015). These proposalsinclude:

● PromotingaccesstobroadbandtechnologiesandInternet.● Helpingfamiliesteachchildrenandyoungpeople.● Promotingcontinuouslearninginschools,kindergartensandhighereducation.● Supporting deployment/enhancement ofWi-Fi access points and education for adults

whoarenolongerinschool.● Promotingbasicliteracy/numeracyskills—thefoundationsofdigitaleducation.

44http://civictech.ca/


February2018


Manyhaverecommendedthatdigitaleducationfocusonarangeoftechnologies, ratherthanmerelylearningapps,whileencouragingcritical,informeduseoftechnology(Peres,2015;IERC,2015),accompaniedbycreativeandproductiveskills.IthasalsobeensuggestedthatstudentsbeintroducedtothreebasicITconcepts:language,informationandalgorithms(Conseilnationaldunumérique,2013).

IntermsofinstallingnewandreinforcingexistingWi-Fiaccesspointsandprovidingeducation,supportnetworkshavebeenproposedthattakeintoaccounttheongoingeducationrequiredbyaconstantlyevolvingenvironment(ConseilNationalduNumérique,2013).Digitaleducationwillalsobenefitfromnotmerelybeingperceivedaswayof“catchingup,”butastraininggearedtothedevelopmentofcreativity,alongwithpersonalandcollectivegrowth.

5.2.2 UnitingPeopleAroundGoals,RatherthanTechnologyInviewofthedigitaldivideandvaryinglevelsofpublicengagement,thesmartcity’scollectivegoalsshouldbesharedtosupportbroad-basedsocialinclusion.Inotherwords,socialinclusionshould not depend solely on digital numeracy, but on the ability to understand, discuss andimprovetheIoTproject,andultimatelyparticipatefullyinit,infullknowledgeofthefacts.

5.2.3 DigitallySupportingthePublic’sPoliticalPower

DanielKaplan(2012)suggestedmakingthecitynotjustadistributorofinformation,butofpower.Thenewpartnershipofcitygovernment,businessandresidentsthatthesmartcityoffersisnotnecessarily limitedto“openingdata,”but to facilitatingpioneeringconceptsthatcan increasesocial harmony.According to Kaplan, there are abundant examples of such ideas: “mobilizingneighborsand shopkeepers tohelp seniors restoreWi-Fi access,where ithasdisappeared, topublicandprivateservices,sharingexpensive,underusedequipment...recyclingandprovidingtransportation for poorly served neighbourhoods.” (Kaplan, 2012). The smart city shouldaccordinglyprovideresourcesforputtingdataintothepropercontextandtoolsfortakingsuchactionsas“creatingprogramminginterfacesoncertainmunicipalapps(feecalculation,mapping,etc.), organizing tools and giving presentations, etc., to facilitate use of these resources byunspecializedplayers; training theseunspecializedplayers tohelp themempowerthemselves;openmeeting,coproductionandmutualassistancefacilities,andshowcasingactivitiessupportedbydigitaltools”(Kaplan,2012).

Others also emphasize the profoundly political role of e-inclusion. To be truly included in apoliticalprojectaimedatachievingsocialharmony,peoplemustnotonlybemembersofsociety,butbeengagedasstakeholdersintermsoftheirsocialbondsandcontributionstoeconomicandculturallife.Whilemanysuchinteractionsarenowdigital,skillsinsupportingthiskindofsocialinclusion are also necessary. This means the technical aspect of the digital world must berepoliticized (Conseil National du Numérique, 2013). In other words, a technological projectshouldnotbeperceivedasnonpartisanandapolitical,butratherascentral tothedemocraticprocess.Suchaprojectshouldalsobeperceivedasembodyingvaluesandcontributingto thereinforcementorreconfigurationofoursociety’spowerrelations.


February2018


PublicparticipationintheIoTprojectisasubjectdiscussedinseveralofthepotentialsolutionspresented in this report,particularly inSection4 (Ethical Issue:Privacy)andSection9 (EthicalIssue:Privacy).

5.2.4 AccessibleandRelevantServicesforAll

Asdescribedabove,digitallymarginalizedgroupshavelesschanceofgettingtheirneedstorankhighly among newly created services (Viitanen and Kingston, 2013). Such needs pertain totechnologyandcontent.Technologically,thevastmajorityofonlineservicesaredesignedforanaudiencewithfunctionalliteracy,numeracyandPSTREskills—ratherthanthoseatthebottomofthelearningcurve.Disenfranchisedgroupsshouldbeincludedatthedesignstagetoincreasethelikelihoodofmeetingtheirspecificneeds.Governmentsalsohavean importantroletoplay indevelopingplatformsthatprovidepublicaccesstodatasets.

5.2.5 AlgorithmicAccountability45

Some writers advocate algorithmic accountability, which means making algorithms auditable(Sandovig,etal.,2014).However,thisapproachhasbeencriticizedbysuchwritersasAntoinetteRouvroyandBernardStiegler (2016),whoexplain thatmakinganalgorithm’sworkings visibledoesnotmeanitcanbeunderstoodorvalidated.

AccordingtoCardon(2015),accountabilitymakesanalgorithmmorerobustbyemphasizingitsdidactic aspect, requiring its creator to make what the algorithm does comprehensible. Analgorithmisaccountablebecauseitprovidesconstantassurancethatitdoeswhatthedesignersaidandthedesignersaidwhatthealgorithmdoes.Theauditmechanismteststhealgorithm’sresults. Itmightbenoted that theTuring Institute’sSuchanaSeth (2017) is currently studyingalgorithmicaccountabilityandthepossibilityofincorporatingethicalcodesinalgorithms.

5.2.6 EthicsinDesign

The IERC (2015),on theotherhand,advocatesethics indesign—makingplayersawareof theprocessthatwillincorporatevaluesandstandardsinalgorithms—tomakethesevalueselectionsvisibleandtransparent.Thisapproachisultimatelyintendedtoincorporaterecommendedvaluesandrightsinalgorithms(IERC,2015).

45Forfurtherinformationonalgorithmicaccountability,seeChristineBalagué,“Plaidoyerpourlaloyautédesalgorithmes,”Éthiquede la rechercheennumérique.Gouvernancedesalgorithms, Feb.2016,Paris,France.2016.<hal-01274665>


February2018


5.2.7 DiscriminationAwareDataMining

These proposed technical solutions are based on Discrimination Aware DataMining (DADM)(Pedreschi,etal.,2009).Theideaistoeliminatetheblackboxinalgorithmsandpreventtheriskofdiscrimination.ForDADM’sauthors(AsmitaKashid,etal.,2017),thisinvolvescreatingadigitalappthatwould: (1)Audit thealgorithmtoquantify itsdiscriminatorypotential. (2)Provideanexplanationoftherulesgoverningdatacollectionandanalysis.(3)Explainthemechanismsusedto mitigate discrimination. Individuals and the community will welcome quantification of analgorithm’sdiscriminatorypotential.

One question arises, however. Can the realignment of relationships between individuals andorganizationsbelimitedtosoftwaretweaks?

5.2.8 RegulatingWhatAlgorithmsCanandCannotDo

Various attempts to establish a legal framework for algorithmic analyses are emerging. TheEuropean Union, for example, has decided to regulate the issue of discriminating algorithms(Goodman and Flaxman, 2016). On April 27, 2016, the European Parliament validated theregulatoryframeworkofthenewEUGeneralDataProtectionRegulationondataproductionandanalysis,toensurepersonalprivacyandpreventalgorithmicbias(EuropeanParliament,2016).

Otherwritersrecommenddefiningtherulesofthegame—whatanalgorithmcanorcannotdo.GoodmanandFlaxman(2016)suggestedsettinguptestingagenciestodetermineifanalgorithmdoeswhat it is said todo. Theyalso recommendeddevelopinga setof rulesapplicable toalldesigners(GoodmanandFlaxman,2016).OnMay30,2017,BenShneiderman46proposedtotheprestigiousAlanTuring Instituteof London thecreationofaNationalAlgorithmSafetyBoard,modelled after transportation boards. The notion of control mechanisms is clearly makingheadway.

46https://www.turing.ac.uk/events/turing-lecture-algorithmic-accountability/


February2018


6 EthicalIssue:SeparationoftheGovernmentandBusinessSpheres

UrbanIoTposesethicalissuespertainingtoseparationofthegovernmentandbusinessspheres.AsdiscussedintheIntroductiontothisreport,theSmartCityProjecthas,fromthestart,beenstronglypromotedbyprivate interests, suchas Internetandcommunications technology (ICT)firms (Carlsson, 2014; Kaplan, 2012; Greenfield, 2014). At the present time, IoT equipment,hardwareandsoftwareforthesmartcityaremostlysuppliedbytheprivatesector,whichhastherequiredknow-howfordevelopingtheseleading-edgedevices.Furthermore,aprojectinvolvingthe release of certain data to the public also implies significant private sector involvement inanalyzingandmobilizingdatausedinthecreationofappsforthepublicandtheprivatesector.Opendata,ofcourse,isalsointendedforusebypublicentities,mediaandthepublic.However,itseemslikelythattheprivatesectorwillbeattheheadofthequeuetoexploitthisnewdatawindfall.

Thestronginfluenceonthesmartcityofprivateplayerscouldunderminetheseparationofthegovernmentandbusinessspheres,posingethicalissues.ÉdithDeleury,formerpresidentoftheCommissionàl’éthiqueensciencesettechnologiesduQuébec(CEST),isconcernedabouthowtheprivatesectorisacquiringgrowingcontrolovermunicipaldataandservices(Deleury,2016).

6.1 PotentialThreatstoSeparationoftheGovernmentandBusinessSpheres

PossibleriskstotheseparationofthegovernmentandbusinessspherespertaintothreeofthefourmainphasesofUrbanIoT:

● IoTinfrastructureplanningandmaintenance.● Dataanalysis.● Opendataandservicesforthepublic.

6.1.1 PotentialThreatstoPlanningandMaintainingIoTInfrastructure

AttheIoTinfrastructureplanningandmaintenancephase,threepossibleriskstoseparationofthegovernmentandbusinesssphereshavebeenidentified:

● Shapingthesmartcityprojectaroundprivateinterests.● Locking-intheprojecttechnologically.47

47Thelinkingofcitiestotechnologyplatformsorvendorsoverlongperiodscreatesmonopolies(Deleury,2016;Kitchin,2014b;Angelidou,2015;Hill2013).


February2018


ShapingtheSmartCityProjectAroundPrivateInterests

RobKitchin(2014b)believessmartcitygovernance isusuallycooptedandexplicitlyshapedbyprivateinterestsfortheirownbenefit(Kitchin,2014b).TechnologyremainsthekeytothesmartcitydesignsandvisionsofsuchfirmsasIBM,CiscoSystems,SiemesAG,Nokia,Veolia,Dassault,GeneralElectric,andPhilips,etc.(Albino,BerardiandDangelico,2015;DouayandHenriot,2016).Smart city projects are themselves characterized as apolitical—as solutions not driven byideological interests, but by “common sense” and optimization objectives (Koolhaas, 2014;Oddoux,2016).

Underthesecircumstances,manyobserversareconcernedthatmunicipalitieswill turntooff-the-shelfproducts,ratherthanengagingindetailedassessmentsofresidentneedsandrequests(Deleury,2016).Kitchin(2016)saysthatprivatebusinessessellcitiessolutionsthatignoretheirhistorical, political, social, territorial and cultural contexts. Evgeny Morozov (2015) calls thisapproachsolutionism,meaning that theprivatesectorprovidesanoverlynarrowdefinitionofsocial problems and does so in terms that will primarily benefit the “solution’s” developers(Morozov,2015b).

Thiswas also thepositionofmunicipal representatives at aHarvard LawSchool internationalworkshop,48whoobservedthatbusinessesthatvendtheirserviceswillprimarilypromotetheirownagendas,ratherthanseekingtoimproveresidentqualityoflife(HarvardLawSchool,etal.,2017).“We’reoftenfindingthatthere'sagap,sometimesquitealargegap,betweenthewaythatvendorsapproachusandthekindsofchallengesthatwewanttotakeon,”theysaid(HarvardLawSchool,etal.,2017).

Locking-IntheProjectTechnologically

Governmentdependenceonalimited(sometimesverylimited)numberofbusinessesactingastheprincipaltechnologyprovidersposesathreattofreepolicy-makingbygovernments,alongwiththeirresilienceandflexibility.Itisrarelyusefulforbusinessestocreateandselltechnologiesthat will be easily compatible with equipment and software supplied by other vendors.Furthermore,themanyupdatesrequiredforproperoperationoftheequipment,thedifficultlyforuserstomakechangesandtechnologicalrigiditybolsterthepowerofprivatecompanies—along with government dependence on them (Deleury, 2016). Some cities, incidentally, areentirelygovernedbyprivatefirms,includingMasdarCityinAbuDhabiandSongdoInternationalBusinessDistrictinSouthKorea(Kaplan,2012).

48Organizedin2016andattendedbyrepresentativesof17cities,mostlyintheUSbutfromothernations,aswell.


February2018


Then there are equipment and service contracts that pose issues of pricing and terms.Manymunicipal stakeholders, for example, point to the sometimes-astronomical prices of variousservices,aswellasthemanyvendorsinterestedinprovidingtheirservicesconditionontherighttoreselldataacquiredthroughtheseservicesortousethedataforcommercialprofiling(personalcommunication).

In this context, government officials quickly become dependent on devices developed (andsometimescontrolled)bybusinesses.Localgovernmentscanquicklybedisempoweredbydigitalexpertsandcompaniesexploitingbigdata(Felli,2015).

6.1.2 PotentialThreatsinDataAnalysis

Twothreatstoseparationofthegovernmentandbusinesssphereshavebeenidentifiedatthisstep:

● Developmentofenhancedservicesforthe“public”controlledbybusiness.● Growingdependenceonsellingdatatogeneraterevenue.

DevelopmentofNon-PublicServiceswithGrowth-GenerationPotential

InthecaseofMontréal,thehopeisthattheprivatesector’smobilizationofopendataandappdevelopment will bring innovative solutions to urban issues. However, since many servicescreatedinthismannerarepaidandcontrolledbyprivatebusinesses,someobservershaveputthepubliconguard.AsMorozov(2015)wrote,passingsomeresponsibilityforservicecreationtobusinessesthatproducepaidappsreducesaccesstocertainservicesintendedforwide-scaleuseandthatserveasadd-onstoservicesgenerallyconsideredtofallwithinthemunicipalsphere.

By taking this approach, cities are shedding their ability to administer or organize services(pertaining,forexample,tourbanmobility),astheywouldprefer(Morozov,2015).Theyalsoriskbecomingdependentonthesenewprivateservices,whichcouldbecomecritical,althoughtheircommercialfunctionsmakenotbealignedwithmunicipalgoals.

Thisprocessofservicecreationandenhancementbyprivatebusinessalsocorrespondswithareduced role for and less involvement by government (Oddoux, 2016). Such a situation canheightentheperceptionthatcitygovernmentisabandoningoneofitsfieldsofjurisdiction.

DependenceonIncomeGeneration

DataiscentraltotheurbanIoTproject.Somedescribedataasaresource,othersasanewformofcurrency(BerthierandKempf,2015).Onethingissure:thereisnowabigmarketforsellingandresellingdata.Providingaccesstodataandleasinginfrastructurecangenerateincomeforthegovernment,throughprofit-sharingorsale(HarvardLawSchool,etal.,2017).


February2018


Somecitieshavedevelopedcivicdataexchangeswherebusinessescanobtainhigh-qualitydata(comprehensive,frequentlyupdated,etc.).Afewcitiesareexploringthepossibilityofreleasinghigh-qualitypublicdatatosuchplatforms,whileprovidingdataofslightlylowerquality49toopendata exchanges. Income generated is then used to fund the smart city project (Harvard LawSchool,etal.,2017,p.10).

Underthesecircumstances,thepressureongovernmentstogeneraterevenueishighandcansteerprojectsawayfromthepublicinterest.


6.2.1 ControllingtheTechnologyandAdministeringPartnershipTerms

Tooffsettheprivatesector’sinfluence,citiesareencouragedtosetclearpartnershipconditionswithsuppliersofthesetechnologiesandcontroltheIoTsystem,asmuchaspossible.Insodoing,citiesshould:

● Establishandcommunicateaclearideaofprojectrequirementsandvalues.● Generatedataonitsown,totheextentpossible.● Retainownershipofthedata.● Encouragecompetitionamongvendors.

6.2.2 Establishing/CommunicatingaClearIdeaofProjectRequirementsandValues

ManymunicipalrepresentativeshaveemphasizedtheimportanceofidentifyingpublicneedsandvaluesintrinsictotheIoTproject,inshapingit.Thisapproachprovidesagovernmentwithcriteriaforguiding thedevelopmentof infrastructure,alongwithdataprocessingandanalysis, in linewiththepublicgood.

Manymunicipalrepresentativesalsounderscoretheimportanceofstayingfocusedonmunicipalneeds,withtechnologynotanendinitselfbutawayofmeetingthecity’sgoals.Theynotetheimportanceoflearningtosaynotovendorsolutionsthatarenottailoredtosuchgoals(HarvardLawSchool,etal.,2017).

Educatingandtrainingvendorsonthistopicisessential(HarvardLawSchool,2017).Somecitiesinvitebusinessestotheirofficestoexplainspecificneedsanddeterminewhichcanbemetbyexistingsolutionsorthroughthedevelopmentofnewproducts.Thisapproachcanturnasalesrelationship into a partnership and expand the range of products or services offered by thevendor.

49 Such lowerqualitydoesnotmeans thedata is poor.Rather, it refers to sounddata that lacks somefeaturesofhigh-qualitydata,suchasfrequentupdates.


February2018


6.2.3 GeneratingandOwningData

Expertshavealsodiscussed the importanceofa citygenerating itsowndataandmaintainingownershipofit.Morozov(2015)saidthatcitiesshouldtrytogeneratethedatatheyneedfortheirownmanagementandthendecideiftheydoordonotwanttopermitprivatefirmstouseitandunderwhat terms.NewYorkandChicago, forexample,are trying to launcha central app fordispatching conventional taxis with the ease of Uber. In addition to contending with Uber’smarketdomination,theprogramwillpreventtripdatafrombecomingacostlycommoditythatcitygovernmentsmustpurchaseathighprices(Morozov,2015).

Box7:CitiesattheWheelofInnovation

Somecitieshaverolledoutappstoinformresidentsofalltransitoptions,fromself-servebikesonthecorner,tominibuseswithitinerariestailoredtopassengerneeds.

Helsinki,inpartnershipwiththeAjelostart-up,createdKutsuplus,acrossbetweenUberandatraditionaltransitsystem.“Passengersrequestashuttleontheirphoneandtheappcalculatesthebestwayofgettingeveryonetotheirdestinations,usingreal-timedata.Italsoprovidesanestimateoftravellingtime,byKutsuplusandotherformsoftransportation”(Morozov,2015b).

Suchaprojectcanonlybesuccessfulifcitieslookbeyondexistingsolutions.ConsideringUbertobetheonlywayofboostingtheefficiencyofpublictransitandreducingtrafficjamsisnotthebeststartingpoint.Morozovalsosuggestedthatstrugglesoverprovidingservicestothepublicwillbewonbythoseowningthedataandthesensorsthatgenerateit.“ByleavingallofthattoUber—orworse,togianthi-techfirmsseekingtocaptureashareofthelucrative“smartcity”market,citiesarepassingupthechancetoconductexperimentsthatwouldenablecommunitiestoorganizetheirtransportationsystemsastheyseefit”(Morozov,2015b).

Inshort,“ItisnotbecomeUbercomesfromCalifornia—aregionfamousforpoorpublictransit—thatweshouldthinkpersonalmotorvehiclesarethefutureoftransportation”(Morozov,2015b).

6.2.4 BreakingupMonopolies

Onewayofensuringgreatercontroloverurbantechnologyistokeepprivatefirmsfromacquiringamonopolyoverit.Thevastnumberandtypesofvendorsgivethecitymorefreedomandletthemmaintainacentralroleasprojectoverseer(Pouilly,2014).Thisapproachalsoletsthecitytakean innovativeapproach toharmonizinghard-andsoftwaresuppliedbydifferentvendorswithintheproject—yieldingafarmoreresilientarchitecture.


February2018


7 EthicalIssue:TransparencyandReliability

Deploying IoT technology in the smart city and analyzing the data it generates raise ethicalquestionsoftransparencyandreliability.TheseissuesapplytothreeofurbanIoT’sfoursteps:

● Datacollectionandstorage.● Dataanalysis.● Opendataandmunicipalservices.


Onethreattoprivacyhasbeenidentifiedatthisstep:

● Datasystemsecurity.

DataSystemSecurity

SystemsecurityisanimportantissueintermsofthecredibilityandreliabilityofasystemlikeIoTinthesmartcity.Section4.4.1discussesthevulnerabilitiesmentionedintheliterature.Intermsof system reliability, it should be noted that highly automated and interconnected systems(particularlyindustrialones)areatgreaterriskofbeinghacked(AmericanInternationalGroup,2016). In the case of a connected city, hacking could immobilize the operating and decision-making processes residents need for crucial services. The crashing or hacking of automatedsystems also raise questions of liability, while underscoring the social, environmental andeconomicimpactofsuchproblems.Inotherwords,thesepotentialissuespertaintodailyliving,asmuchastopublicsafety.

7.1.2 PotentialThreatsofDataAnalysis

Onethreattoprivacyhasbeenidentified:

● Lackoftransparencyoftechnologicalsystemsandanalyses.

OpacityofSystemsandAnalysesUsed

The complexity and lack of transparency of current data analyses and systems directly affectpeople’srighttoknowwhatwillbedonewiththeirdata(EuropeanParliament,2015).Thepublichasonlyafoggynotionofthekindsofdatacollected,howthisisdone,whyitisanalyzedandthekind of technology concerned. Furthermore, big data analyses often include data analysis forpurposesotherthanthosespecifiedduringcollection(EuropeanParliament,2015).


February2018


Someobservers, however, say that extensive communicationwith the public on the analysesperformed and the system architecture would create additional security risks for systems orindividuals trying to game the system by such means as generating false data through theiractivitiesorbycomputer.Thereisafinebalance,inotherwords,betweenthedesiretoopendataandtheneedtorefrainfromsayinganything,particularlyinviewofpublicsecurity(RichardsandKing,2014).

7.1.3 PotentialThreatsofOpenDataandServicesforthePublic

Onethreattoreliabilityhasbeenidentifiedatthisstep:

● Qualityofgeneratedandopendata.

QualityofGeneratedandOpenData

Datamustbeofhighqualitytoensurethat itcanconfidentlybeusedindecision-making(Lee,2017).However,opendatareleasedbygovernmentagenciesisnotperceivedtobehighinquality(personal communication, 2017). The work of McArdle and Kitchin (2014), drawing on theirexperience as developers of applications employing urban data, highlights the difficulty inevaluatingdataaccuracywithoutqualityreportsfromdataproviders(McArdleandKitchin,2014,p.1).

Theauthorsnotedthatopen-dataportalsusuallydonotincludeenoughmetadatatoletusersassessdataquality (p. 9).A studyof dataportals in London, Paris andDublin reveals that noinformationongeneralor specificqualitymeasuresaccompanies theirdata.While somedatatracking elements, such as the data’s age and its supplier’s name, is shared, the process oftransformingthedatafromarawcommoditytoafinishedproductisnotdescribed(McArdleandKitchin,2014,p.9).Moreoftenthannot,opendataisprovided“asis,”withnoassuranceastoitsaccuracy,continuityoftraceability.

McArdleandKitchinwarnthatifgovernmentagenciesfailtodealwithdataaccuracyissues,open-data portals may be perceived as potential “dumping grounds for unreliable, unaudited andpoorly preserved data” (McArdle and Kitchin, 2014, p. 9). The authors acknowledge that theultimateproblemisrarelylackofinterestbyofficials,butinadequateresources.


February2018



CrowdsourcingtoIncreaseMetadata

McArdleandKitchin(2014)arguethat,intheabsenceofdataprovidersusingmetadataandusermanuals to document their data quality, open-data portals should employ a crowdsourcingmechanismtogenerateandrecorduserfeedbackandtweakstoimprovethequalityofdatafromurbansourcesandopengovernmentportals(McArdleandKitchin,2014,p.1).Thiswouldpermitsubsequentworkbyothers,witherrorsdetected,andproblemswithorusesofthedatasharedinthesamemanneraswithvolunteergeographicinformation(VGI)systems.

MonitoringDataQualitywithMetricsandStandards

Avarietyofguidelinesandmeasureshavebeenproposedtodefinedataqualityrulestoobserve(Batini,etal.,2009).ISOdataqualitystandardshavebeendeveloped,includingISO19115-1:2014(Geographic information—Metadata) and ISO 19157:2013 (Geographic information— Dataquality).McArdleandKitchinalsorefertotheworkoftheInternationalCartographicAssociation,which has identified seven metrics associated with the accuracy of spatial data.50 They alsomentionedworkontransportationbythescientificcommunity,whichfoundwaysofdisclosingtrafficdataquality(Turner2002inMcArdleandKitchin,2015).

TheUSEnvironmentalProtectionAgency(USEPA)hasselectedfourquestionstoanswerwhenpublishingenvironmentaldata,topermituserstoassessthequalityofthisdataanddetermineifitcorrespondswithitsintendedpurpose(USEPA2006).ThesequestionsappearinAppendixD.

Then,thereistheOpenDataInstitute,whichcreatedtheOpenDataCertificatetoenhancethecredibility of data supplied by data providers. Suppliers auto-certify by answering a set ofquestionsabouttheirdata.Adescriptionoftheirqualitycontrolprocedureissubmittedwiththedatatoobtaincertification(ODI,2015).TheEUINSPIREDirectivealsorequiresthatgeographicandtrackingmetadatabeprovidedalongwiththedataitself(Inspire,2015).

7.2.1 BuildingTrustThroughTransparency

Transparencycanplayakeyroleinbuildingtrustbetweenresidentsandgovernmentagencies.Itcan also be instrumental in preventing abuses of institutional power. Transparency permitshealthy checks and balanceswithin government and between the government and governed(RichardsandKing,2014).Somestakeholdersandexpertshaveidentifiedwaysofbuildingtrustthroughtransparency,including:

50“Lineage,positionalaccuracy,attributeaccuracy,completeness,logicalconsistency,semanticaccuracy,temporaldata”(GuptillandMorrisson,1995,inMcArdleandKitchin2015).


February2018


● Gradualdeployment:startslowly,limitingdatacollection,andtakethetimetoexplainnewprojectstothepublic.

● Beclearonwhatdataisnotcollected.● Clearlyexplainkeyanalyticoperations,goalsandsubsequentmanipulationsofthedata

collected,andidentifyanythirdparties involved,whileallowingforfutureadaptabilityandevolution.

● Clearlyidentifywhatdataisconsideredsensitive(personal,geolocation)andhowitwillbehandled.

● Letpeopleknowwhatthecityhasalreadydoneorimplemented(HarvardLawSchool,etal.,2017;EuropeanParliament,2015).

Availablechannelsofcommunicationandinformationtechnologiescanbeusedtoprovidesuchinformation.However,one-on-onehumaninteractionisfundamentaltobuildingtrustbetweenthepublicandgovernmentagenciesandshouldnotbeneglected.Accordingtovariousmunicipalrepresentatives,suchinteractionisessentialtothesuccessoflong-terminitiatives(HarvardLawSchool,etal.,2017,p.12).

Section4(EthicalIssue:Privacy)alsoidentifiedtransparencyasapossiblesolution.


February2018


8 EthicalIssue:Freedom

ThedeploymentofurbanIoTposesethicalissuespertainingtopersonalfreedom,autonomyandself-determination.TheseissuesareassociatedwithtwoofthefourphasesofurbanIoT:

● Datacollectionandstorage.● Dataanalysis.


Twothreatstofreedomhavebeenidentifiedatthisstep:

● Actualorperceiveduniversalsurveillance.● Greaterdataveillanceduetotheenhancementofavailabledata.

ActualorPerceivedUniversalSurveillance

DeployinganInternetofThingsisaccompaniedbytransparencyandthereleaseofdatathatcanempowerthepublicthroughsuchmeansaspermittingexamineitsenvironmentandtheactionsof city hall more closely, as well as by developing local solutions to problems. However, anenvironmentcrammedwithdatasensorscanhaveamuzzlingeffectonthepublic,whosewordsandactionsareconstantlybeingconvertedtodigitalcode.

According to psychological and psychoanalytic thinking, constant subjection to observation,scrutinyandexaminationcantriggera“suspensionofidentity”(Birman,2011,40).Peoplewhoareconstantlyunderwatchfeellikeprisonersofanomnipresent“gaze”andseektobreakfreeofit. It has been shown, for example, that surveillance has an impact on self-censorship andcensorshipofminorityopinions(Richards,2013;Stoycheff,2016).Suchsurveillanceconsequentlyhasan impactonfreedomofexpressionandthought—coreprinciplesofa freesocietyandofmosttheoriesofdemocraticpoliticalfreedom(Richards,2013,1951).


February2018


Box8:ThePanopticCity

Many writers recount Bentham’s Panopticon metaphor to describe the kinds of problemsresulting fromubiquitous surveillance.ThePanopticon is a circularprisonwithawatchtowerwhereaguardmaybepresent,butwhocannotbeseenfromtheoutside.Thetowerfacesmanycellsofprisoners.Theinsideofeverycellcanbeviewedfromthistower.ThemaineffectofthePanopticonistomakeinmatesfeelasiftheyareconstantlyundersurveillance.Thisensuresthatpowerover them isautomaticallymaintained.Theeffectsof such surveillanceare relentless,even if suchmonitoring isnot.Perfectpower tends tomake itsapplicationunnecessary . . .”(Foucault,1975,p.234).Thepower(watchtower)isalwaysvisible,butprisonerscannotbesureifanyoneisactuallywatchingthem,althoughtheyknowtheycouldbeunderobservationatanymoment. The same inductive logic applies to connected objects. Their sensors reap vastquantitiesofdata,butpeopledonotknowinadvancewhatwillbeseen.RobKitchinwritesthatdeployingurbanIoTcouldpotentiallycreateapanopticcity(Kitchin,2014)51.

IncreasingDataveillancetoEnhanceAvailableData

“Dataveillance”isaformofsurveillanceinvolvingthecollection,sortingandaggregationofdatasets to identify, track,monitor, predict and constrain personal behaviour (Clarke, 1988; Raley2013;Kitchin,2016).Suchsurveillanceispossiblethankstovastquantitiesofavailabledataandcontemporary analytical techniques.With theextradata it collects, IoT canplay a key role inbolsteringdataveillance.

51Somewritersincludeintheconceptofsurveillanceintentionalorunintentionalsubveillanceononeselfbyothers,using“smart”technologies.PleaseseeAppendixEformoreinformationonthistopic.


February2018


Dataveillanceinvolvesreal-timedataharvestingandstorageofvastquantitiesofbulkdigitaldatafrom different sources. Harvested data may be generated by personal digital devices (smartphones,computers,wearablecomputers,creditcards,RFIDtags,etc.),lifelogging,52digitalsocialnetworksandotherconnectedapps,aswellasinteractionswithdifferentsensorsthroughoutthecity.Collecteddataisaresourcethatcanbesubsequentlyenhancedthroughaggregationwithotherdata.AccordingtoRouvrayandBerns(2013), thiskindofconstantdatacollectionotherthanforpredictivepurposesandwithoutanyspecificpurposeisunprecedentedinthehistoryofprofiling (Rouvroy andBerns, 2013).Analysis, primarily of correlationsbetweendata, follows,withdigitalduplicates(HaggertyandEricson,2006;RouvroyandBerns,2010)generated,alongwithvariouspredictionsofpersonalbehaviour.ThisprocessclearlydemonstratesthatthelargerdataquantitiesgeneratedbyurbanIoTwouldimplymoreeffectiveandpreciseanalysesforuseinmonitoringindividuals.

In addition to all of thepreviouslymentioned issuesof self-censorship, disempowerment andsuppression of minority viewpoints, dataveillance, and in particular, preemptive prediction,53posechallengestothepresumptionofinnocencerule.Lackoftransparencywithrespecttothedata,accompaniedbyalgorithmicreasoning,aswellasdecision-makingbasedontheexpectationof bad behaviour, may contravene the principle that everyone is presumed innocent unlessconvicted.Wemustnowbesuspiciousinadvance,contrarytothedoctrinethatthatweshould“first, trust thewordofothers, thendoubt if thereare strong reasons fordoing so” (Ricœur,2000),inlinewiththeconceptofsocialhabitusandrulesofprudence.Thisstanceunderminesoursenseofcommunityandthesocialbond.Omnipresent,constantlyaccessibledatabecomesa“proof”ofanythingandmayreplacealltestimony,orevendiscussion,dealingaheavyblowtosocialharmonyandincreasing“socialanxiety.”

8.1.2 PotentialThreatsinDataAnalysis

Threethreatstofreedomhavebeenidentifiedatthisstep:

● Prescriptiveanalyses,whichguidepersonalchoices.● Predictiveanalyses,whichdetermineindividualaccesstoopportunities.● Profiling,whichinhibitspeople’snaturalabilitytotransformthemselves.

52Lifeloggingistheaccumulationofquantitativepersonaldataconcerningdifferentaspectsofaperson’slife(health,relationships,sportsperformance,etc.).PleaseseeAppendixEformoreinformationonthistopic.53Pleaserefertothenextsectiononforecastanalysisforadditionalinformationonpreemptiveprediction.


February2018


Section5.1.2,above,identifiespossiblethreatsassociatedwithdataanalysesusedforforensicprofiling.

PredictiveAnalysesthatGuidePersonalChoices

KerrandEarle(2013)proposedageneraltypologyofanticipatoryalgorithmsthatpredicthumanbehaviour.TwoofthethreetypesmayberelevanttotheUrbanIoTproject.54Theyare:

● Preferentialprediction.● Preemptiveprediction.

Preferentialpredictionconcernsopportunitiesandchoicesthatpeopleshouldhave.Itemploysmachine-learning algorithms to predict which information and products are likely to interestconsumers.Basedontheirdigitalprofilesandactivityhistories,peopleareofferedchoicesthatpresumably“matchtheseprofiles.”Clearly,themoststrikingexamplefromtheprivatesectorisAmazon’s 2013 recommendation technology patent filing: based on digital profiles, thetechnologymakesitpossibletoshipmerchandisetocustomersbeforetheyhaveeventhoughtofbuying it (Bensinger, 2014). These strategies are certainly aimed at making information andproductsearchesmoreefficientforpeople,whileatthesametimetheyservetochannelandlimitpersonal choices and, generally, to offermore or less the same kinds of content that peoplepreviously selected (Taylor, 2014). The question arises: howmuch autonomy do people havewhenchoicesabouttheirownlivesareconstantlyinferredbyothers?

PredictiveAnalysesDeterminingPersonalAccesstoOpportunities

Preemptiveorprescriptivepredictionisintendedtofacilitatedecisionsonprovidingaperson(orgroup)withaccesstoservicesoropportunities.Ifalgorithm-basedpredictiveanalysesandprofilescanbeconsultedbythirdpartiesoraresubjecttoautomateddecision-making,theycanhaveamajorimpactonfindingwork(dependingonemployabilityscore),gettingcredit(dependingoncreditrating),obtaininghealthorcarinsurance(dependingonriskrating),beingadmittedtoaschool,etc.Peoplemaynotbeabletoreceivecertainservices,basedontheirdigitalduplicates.55Suchtechniques,ofcourse,alsoapplytoprofilingforpublicsafetypurposes.

54Thethirdtypeisconsequentialprediction,whichseekstopredicttheconsequencesofpeople’sactions,sothattheycanbeencouragedtofollowrules“toimprove”theirconduct.PleaseseeAppendixEformoreinformationonthistopic.55Observersnote,however,thatthe“knowledge”producedbyalgorithmsisinferredthroughcorrelations(andnottheotherwayaround),dependingonselectedcriteria.Thisprocesscanbeperceivedasbeingmoreobjectivebecauseitismachinebased,butitremainssubjecttoincorrectinterpretation.Itcouldevenconfusecorelationshipswithcausality—accompaniedbyasevereimpactwhenlegaldecisionsarebasedonsuchdata.


February2018


Box9:InabilitytoDisputeDecisionsBasedonPreemptive/PrescriptivePrediction

Preemptive or prescriptive prediction also raises the issue of people’s ability to disputealgorithm-baseddecisionsabout them.Peopleusuallyhavenowayofdetermining thepathstakenbycollecteddataortheinformationgeneratedbyalgorithms.Severalobserversdescribealossofcontrolduetothefactpeopledonotknowwhenorhowtheirprofilesareused(Kitchin,2016). The relatively unintelligible nature (even for programmers) of algorithmic reasoningmakes it farmore difficult to appeal a decision (Kerr and Earle, 2013). Giving an account ofyourselforyouractionswhenfacedwithalgorithmicdecisionsandassumptionsisafundamentalchallenge.Therighttoexplainyourself,particularlyinalegalsituation,mustnotbeeclipsedbycorrelationalstatisticsusedas“objective”evidence.

TheimportanceofpotentialremedieshasalsobeendiscussedindetailintheSection4(EthicalIssue:Privacy)andSection5(EthicalIssue:SocialInclusion).

ProfilingInhibitsPeople’sNaturalAbilitytoTransformThemselves

Profiles built around a plurality of data values obtained fromwithin a person’s environmentcreates a lasting portrait of that person.56 As Rouvroy wrote, though, one key to personalempowermentis“theopportunitytoseeyourlifenotasaconfirmationorrepetitionofyourdatatracks,butasachancetochangedirection,explorenewlifestylesandwaysofliving—inshort,todowhatothers,andevenyou,donotexpect”(Rouvroy,2008).Withlarge-scaledeploymentofalgorithmicprofiling,itwillbecomeincreasinglydifficultforpeopletostartover.AsMannermaawrote,theubiquitoussurveillancesocietyneverforgets(Mannermaa,2007,112).

56Akindof“totaldigitalmemory.”


February2018



Thissectionoutlinesa fewadditional tentativesolutionstoethical issuesofpersonal freedomthatwerenotpreviouslydiscussedinthisreport.

8.2.1 RespectingPrivacyRights

Section4,above,discussedmanyrecommendationsonthistopic.Theliteratureontheconceptoffreedomparticularlyemphasizesthefollowingpoints:

● Minimizingcollecteddatabylimitingcollectiontodatacorrespondingwithpubliclystatedgoals.

● Tellingthepublichowthedatacollectedwillbeused.● Adoptingthepolicyofdeletingunuseddata.● Adoptingapolicyondatareuse–otherthaninexceptionalcases,datashouldnotbeused

by any parties that have not been described and stated in advance (Abiteboul andFroidevaux,2016).

Severalof thesepoints—especially thoseconcerningminimizingdatacollectionandrestrictingdatareuse—runcountertotheIoTproject’sgoalsofcollectingbigdatatoencourageextensivereuse.Cityagenciesmustlearnhowtodealwithsuchconflictsovercomingmonthsandyears.

8.2.2 BuildingaRegulatoryFrameworkAroundCertainKindsofAlgorithmicPredictionsandDecidingWhichShouldBeBanned

Thesectiononsocialinclusion(5.2.7),above,discussesthisoption,whichalsoappliestotheissueoffreedom.

8.2.3 DevelopingPoliciesontheRighttoCommentonandContestData

This means formalizing the public’s right to dispute algorithmic decisions that concern thempersonally.Rouvroyrecommendedthat,inadditiontothisrequirement,theburdenofproofbereversed, making a person or institution using algorithms responsible for demonstrating, forexample, that there was no discrimination. These means offering a rationale for algorithmicdecisionsthatostensiblyaffectaparticularindividual(Rouvroy,2016,49).Thisapproachenableseachofthepartiestocommunicate,speakfreelyandgivetheirreasons.Sucharequirementtoprovideexplanationsdoesnotrelieveinstitutionsusingalgorithmsoftheirresponsibilities.Thesectiononprivacyissues(4.5.5)alsodiscussesaccesstoredressmechanismsasanoption(4.5.5).


February2018


8.2.4 DevelopingPoliciesontheRighttobeDisconnected

Thisrightservestoreassurepeoplethatthereisnodigitalprofileonthemorthattheirpersonaldataisonlypartial.“Disconnected”alsomeans“non-digital”andletspeoplegoonwiththeirlivesoutsidethedigitalrealm.Payingtuitiononline,votingonlineandmakingdoctor’sappointmentshavebecome sopredominant it is sometimedifficult to do somethingwithout using a digitalplatform.Initiativesaimedatmaintainingface-to-faceservicesareveryusefulinhelpingpeoplegodigital—ornot.

8.2.5 DevelopingPoliciesonaRighttobeForgotten

TherighttobeforgottenisarecentlegalprinciplethatdoesnotapplyinCanada.57Itisgenerallydefined as “the right to remove or inhibit access to accurate, inaccurate and obsolete onlineinformation—whetheraccurate,inaccurateorobsolete—aboutaperson’spast,sotheycanberemovedthecollectivememoryandforgotten”(Lecomte,2017).Incurrentjurisdictions,suchasEurope in particular, it is the right to delete and correct personal—and generally publiclyaccessible—digitaldata.Applyingarighttobeforgottentobigdata—atechnologyfallingintothegreyspacebetweenprivateandpubliclife—isarealbutimportantchallenge,giventheexistinglegalvacuumonthistopic.

Box10:RighttoEraseData

Article 17 of the European Regulation of 27 April 2016 provides the right to the erasure ofpersonaldata.Article65addsthat“Adatasubjectshouldhavetherighttohavepersonaldataconcerninghimorher rectifiedanda ‘right tobe forgotten’ . . . Inparticular,adata subjectshouldhavetherighttohavehisorherpersonaldataerasedandnolongerprocessedwherethepersonaldataarenolongernecessaryinrelationtothepurposesforwhichtheyarecollectedorotherwiseprocessed,whereadatasubjecthaswithdrawnhisorherconsentorobjectstotheprocessingofpersonaldataconcerninghimorher...”(EuropeanParliament,2016).

57However,theOfficeofthePrivacyCommissionerofCanadahasbeenexaminingtherighttobeforgottensince2016(Lacombe,2017).


February2018


8.3 IsAlgorithmicBiasaNecessaryEvilinEnsuringPublicSafety?

OnNovember25,2016,PiotrSmolarpublishedaninterestingarticleinLeMondeentitled:“SecretAlgorithms for Israeli Surveillance in Jordan.” Wemention thisherebecause it illustrates thecomplexnatureofapplyingpotentiallydiscriminatingalgorithmstopublicsecuritypurposes.

PiotrSmolarexplainedthattheIsraelDefenceForces(IDF)useprofilingappstodetectsuspectprofilesonsocialnetworks.“Thispreventedhundredsofattacks,evenifwecannotbecertainthatevery person identified would have perpetrated one imminently” (Smolar, 2016). Machine-learningalgorithmsareaccordinglydesigned to forecastwhen someonewill go fromword todeed,permittinghisorher apprehensionprior to the crime,evenwhen there isnoobjectiveevidencetosuggestaterroristattackwasabouttooccur.GuillaumeChampeau(2016),anotherjournalistwhowroteonasimilartopic,said:

“According toan Israeliofficialquestionedbyapressagency, theyconstantlydrawupsuspectprofiles,notonlyusingmetadataonaperson’scommunicationsandhabits,butsocialnetworkingexchanges,ofcourse.”(Champeau,2016)

Questions of privacy rights and freedomhold little sway in this security scheme.Willmakingexceptional surveillance measures the new normal give governments powers that, indemocracies,fallunderthejusticesystem?


February2018


9 TransformingGovernanceModes

TheliteratureonsocialissuesassociatedwiththesmartcityandIoTsuggeststhatIoTdeploymentwillcausebasicchangesingovernance.

Whilenotexactly“ethicalissues,”suchfactorsremainimportantinacitygovernment’sday-to-dayactivitiesandinteractionswithresidents.SuchchangesshouldaccordinglybeconsideredinanyassessmentofanIoTproject’ssocialacceptability.

9.1 InternetofThings:FromDriveroftoFactorinChange

Thedigitaleraishereanditishightimetofocusongovernance!AccordingtotheGovernanceWorkingGroupoftheInternationalInstituteofAdministrativeSciences(1996),governanceisaprocess that enables social forces to exert power and authority, define policies andmake orinfluence decisions on public life, as well as social and economic development. Governanceimpliesinteractionbetweenformalinstitutionsandcivilsociety.

The document Infrastructures et villes intelligentes by the Commission de la science et latechniqueauservicedudéveloppement(ConseilÉconomiqueetSocial,2016)doesnotdescribethesmartcityasaresourceforexecutinganurbanplan,butasameansofrenovatingmunicipalgovernance. Studies on digital governance reveal changes in governance itself. IoT, in otherwords,isnotmerelyatechnicalinnovationtobeconsideredforadoptionbyMontréal,butthesource of new governance issueswithwhich they citymust contend.We shall discuss this ingreaterdepth,below.

9.1.1 FactorsofTransformationinMunicipalGovernance

AlloftheidentifiedfactorspertaintothetechnicalorientationtypicalofIoT-basedgovernance,wheretechnologyplaysadominantroleingovernanceanddecision-making.Thistechnologicalfocusentailsadepolitizationoftheconcernsasmartcityisexpectedtoaddress.ThevasturbanIoTproject ispresentedprimarilyasatechnological,apoliticaleffortthatmakes“goodsense.”Thisapproachminimizesdebatesoverprioritiesandproposedsolutionsand,insodoing,reducesthechancesofsocialopposition(DouayandHenriot,2016).

Thekeyfactorsinvolvedintransformingurbangovernanceareallpresentthroughoutthedataanalysisphase.Theyare:


February2018


● Decisionsaimedatoptimization,notsocialoptimumsorbasiccauses.● Dilutionorlossofdecision-makingauthority.● Deterministicrepresentationsoftheworld.● Feweroptions.

DecisionsAimedatOptimization,NotSocialOptimumsorBasicCauses

Akeyfunctionofmachine-learningalgorithmsistodriveoptimization.Onceagain,though,thisidealmustbedefinedand,ifnecessary,rejectedinfavourofalessefficientorcost-effectivebut“smarter”solutionthatactuallysolvesthesocialprobleminquestionandnotjusttheparticularcalculationthealgorithmisdesignedtomake.Achievingsuchoptimizationisatechnicalmatter.Itisimportanttoknowhowtobenefitfromthisoptimization,butthisisnotthesameasasocialoptimum, which results from compromise. Algorithm optimization seeks to achieve multiple(technologically,economic,political,ecological,etc.)optimums,which,whengroupedtogether,can be weighed, discussed and expressed. Strictly speaking, societies do not “function” (G.Canguilhem).

AsKitchinnoted:“Ontheirown,technicalsolutionscannotaddressbasicproblemsbecausetheydonotdealwiththeirrootcauses.Thesesolutionsshould,instead,makeiteasiertodealwithmanifestationsofsuchproblems”(Kitchin,2014b).

DilutionorLossofDecision-MakingAuthority

Twopointsonthistopicbearmention.First,algorithmsgenerateoutputexpressedasdecisions,such as categories and recommendations. Potential human-machine interactions in decision-making fall within a continuum ranging from mere consultation to total delegation ofresponsibilitytotheseanalyticalsystems.Wearewitnessingadilutionorlossofdecision-makingaccountabilitywhere algorithmoptimization takes the formof decisions and actions,withoutpeoplebeingabletoexplainthereasonsforresults,evenafterthefact.

DeterministicViewoftheWorld

In seeking predictability of human and non-human phenomena, people increasingly turn to“stable”correlationsbasedonmachinelearningandalgorithmanalysis.Themassdatausedindifferentformats,fromdifferentsourcesandinastronomicalquantities,isencouraginghumanstoreducesituationstodeterministicmodels(Floridi,2012).Doingsocreatesasituationinwhichstructures are locked into the past. Predictability transforms symbolic representations of thepresentintoanewcategory,knownas“Permanence.”Believingthatthepresentisbetterthantheunknowniscalled“conservatism.”However,weknowthatlifeischaracterizedbyitsmanynewandtotallyunpredictable—butfascinating—changesoftrajectory(Heisenberg,1990).


February2018


Feweroptions

Inhisstudyofthesmartcity,Kitchinnotedthatitsformofgovernancepresupposesthatallofitsdimensionscanbemeasured,monitoredandhandledinthesamemannerastechnicalissues.Heemphasizedthatthisapproachcanshrinktherangeofoptionsandanalysesconsidered,inlinewiththedataavailable.Theanalytic“focus”isconfinedtorawdatathatcanbetransformedintoeasilymanipulateddata(Kitchin,2014b).


9.2.1 MunicipalAssumptionofFullAccountabilityforProject

Decisionsbasedonalgorithmanalysisshouldhelpidentifykeymomentsandfactorsinmunicipaldecision-making,sothecitycanassumefullaccountabilityforthetechnologicalsystemdeployed.Such decision-making assumes that employees are trained in-house to discuss delegations totechnology,toavoidthebureaucratictrapofreplacing“it’snotme,it’sthesystem,”with“it’snotme, it’s the algorithm.” If the government primarily assumes the role of delegating, withoutsupervising,itsauthoritywillbediminishedanditwillbeunabletoaccountforitsownactions.

This is whywe should envision our relationshipswith technology in terms of a technologicalculture that enables us to think of technological devices as part of society—and not merelyinstrumentsthatautomaticallytakeoverourdecision-making.Ahealthygovernmentassumesitsfullroleasprimarydecision-makerandadministrator,whilegivingcivilsocietyfullreigntocriticizeandreclaimdecisionsthathavebeenmadeandpreferredmanagementtechniques(folksonomy,collectivemapping,etc.).

9.2.2 PromotingPublicParticipationintheProject

Wemustbeginbyconsideringhowresidentscanbepartofthisproject,whilealsofocusingonthe development of infrastructure and smart services. Of course, it would be necessary toeducate, train and prepare these stakeholders properly for working with machine-learningalgorithms. The smart city will also—and perhaps primarily—involve members of the publicunabletochoosebetweengoodandbadusesofalgorithms.Thismeansthatthebiggestpublicinvestmentsshouldbeineducation—evenmorethanininfrastructureandnetworks.

In other words, in addition to ensuring optimal management of workflows and smartinfrastructure, themosturgent issue isdetermining the role thepublic is expected toplayasstakeholdersandsubjectsofaproject thatwillgrow—rather thanshrink—theirautonomy (asdefined by Malherbe, 2000). Technological systems must be built around social initiatives.Otherwise,infrastructurespendingmayamounttolittlemorethanconspicuousconsumptionorshort-termurbanwindowdressingforinvestors.


February2018


9.3 DISCUSSION:TheProperRolesforAutomationandAlgorithmsinSociety

It is odd to instinctively label technological systems as “smart” when they seem to performactivitiesorfunctionsautomaticallythatwewerepreviouslyaccustomedtodoingonourown.However,thetermautomatedtypicallyinferslowintelligenceorlittleimagination.Infact,thetermautomationisnotsuitedfordesigningandoptimizingpublicpoliciesandcitymanagement.

Theautomationparadigmis,infact,sooutmoded,thatmachine-learnedoutputcannotalwaysbeanalyzedbytheengineerswhodesignedit.Ithasbecomeincreasinglydifficultforengineerstotracethealgorithmicstepsleadingtoaresult.Wecannolongerrequirethatthesealgorithmsbebothautomatedandinnovativeorautonomous.Intermsofpolicy,theissueiswhatweaskourmachinestodo.

Automation isactuallyavery lowbar,whichsaysmoreaboutour lackoftechnicalknowledge(being able to run a machine properly and contribute actively to its redesign) than thetechnologicalsystem’sintrinsicqualityorperformance(G.Simondon;G.Canguilhem).Ifwereallywanttoliveinsmartercities,withalgorithmicsystemsembeddedwithintheurbanfabric,suchsystemsmustbebetterintegratedwithourvaluesandethics.Thismeansthinkingofalgorithmsasan integralpartofthegovernanceandmanagementprocessesandassocialactorsofferingcertain benefits—and drawbacks (H. Collins). Viewing algorithms in terms of their integrationwithin the social fabric means employing them appropriately, based on our knowledge of asituation, rather than falling into the trapofuniversalautomation,basedonaconviction thatautomatedinstitutionsandactivitiesarebetter,moreefficientandmorecosteffective.


February2018


10 SocialAcceptabilityandIoT

10.1 DefinitionofSocialAcceptability

Québec’sFrench-speakingsociologists58whoarestudying,usinganddevelopingtheconceptofsocial acceptability in Canada share two points of agreement: (1) They recognize a multi-stakeholderconsensusthatthisconceptisusefulinidentifyingandunderstandingthewiderangeofpositionsonanurbandevelopmentproject.(2)Theybelievetheconceptisambiguous,unclearandevendubious.Thisisbecause,despitetheshortcomingsofsocialacceptabilityasananalyticalcategory,itcanbeconsideredacriticalgrouping,whicheachstakeholderinjectswithmeaning,dependingonthatindividual’srelationshiptotheproject.Themeaningsassignedtothisconceptoverthecourseofaprojectbecomequalityindexesofthedeliberative/consultativemechanismsemployed.Consequently,theconceptofsocialacceptabilityseemstopromotethecreationofamulti-stakeholder,multi-levelgovernancesysteminthecaseofMontréal’sIoTproject.

Withthisinmind,wemustassesstherelevanceofthevariousdefinitionsoffered.Thisworkhasbeendone inpartbyPierreBatelier (2015).Someof themost-citeddefinitions in theQuébecliterature appear in Appendix G. As an analytical starting point, however,we shall adopt thedefinitionofsocialacceptabilityproposedbyCorinneGendron(2014),whichhasgeneratedsomeconsensusandseemsusefulinguidingourdiscussion:

“Thepublicsentimentthataplanordecisionresultingfromcollectivewisdomisbetterthantheknownalternatives,includingthestatusquo.”(Gendron,2014,p.124).

Thisdefinition,adaptedfromthatofBrunson,etal.(1996),viewssocialacceptabilityastheresultof a collectivedecision, involving informed choices,with a particularly goodunderstandingofpotential opportunities and benefits, as well as risks. It is also apparent that this collectivejudgmentisbasedoncommunityorgroupvalues,whichareneitheruniformineverysituationorsetinstone.Thisisbecauseopinionscanchangeovertime,inaccordancewitheventsandthepublic’sgrowinginterestinorpolarizationaroundatopicthatgenerallytriggersopposition,andultimately,increasedquantitiesofaccessibleinformation.

58MarieJoséeFortin,CorinneGendronandPierreBatelier,mentioningonlythebest-knownmembersofthisgroup.


February2018


10.2 SocialAcceptabilityoftheUrbanIoTProject

10.2.1 SocialAcceptabilityoftheSmartCityandSocialFactorsintheSmartCityWe have, to date, only found one bibliographic reference, taken from an academic journal,entitled “Smart City concepts: from perception to acceptability” (Schelings and Elsen, 2017).Other writers mentioned below discuss the smart city’s social dimension, but not its socialacceptability.ThisinitialreferenceisdrawnfromtheworkofBelgianresearchersSchelingsandElsenattheUniversitédeLiège,whodevelopedaquestionnaireusedatthreeseparatesmartcityeventsinBelgiumtoassessthesmartcity’ssocialacceptability.Theirquestionnairewascompletedby125participantsintheseevents(attendedbyatotal625people),whoweremembersofthepublicorprofessionalstakeholdersinthesmartcityproject(government officials and service providers). All these people can be considered not justinterestedin,butknowledgeableaboutthesmartcity(SchelingsandElsen,2017),duetotheirpresenceattheseevents.Thequestionnairecovereddifferentfictionalsmartcityscenariosanddidnotpertaintoanyparticularcity.Thisstudy,whichisworthreading,demonstratesthatthemajorconcernofthosesurveyedwasthe “personal,” far more than the economic or environmental, aspect. The top concern waspersonalprivacy.Publicparticipationingovernanceofthesmartcityalsoappearsasathemethatisimportant—butfarlesssothanpersonaldataprotection.

“‘Privatedata’neverthelessemergeasonekeyaspectparticipantswouldbereluctanttoshare,whichunderlinesthedelicatebalanceonehastoreachbetweencollecting largeamountofdata(essentialtonurtureSmartCityinitiatives)andinsuringend-users’privacyandanonymity”(SchelingsandElsen,2017,p.3).

Itisworthnotingthatthisinitialstudyonperceptionsandsocialacceptabilityofthesmartcitywasconductedusingapsychosocialapproachamongasamplegroup.Forthisreason,thestudybenefitsfromconsideringtheconcernsofmanypeopleandnotjustonewriter.Kornberger,etal.(2017),applyingasimilarmethodologicalapproach,predicteda“shock”whenthebureaucracyisconfrontedwithanewopendatapolicyinrollingoutasmartcityprojectforVienna.Theauthorsexpectedaclashofvalueswithincitygovernment,whichhadlittleexperienceinortalentforsharingitsdataanddiscussingdecisionswiththepopulace.Inthecaseunderstudy,theresearchersfoundthatofficialconcernsfocusedontwoareas:(1)Officials’perceptionsoftheiraccountabilityinanopen-datasystem,wheredataisavailabletoeveryoneforawidevarietyofusesoutsidetheircontrol.(2)Transparency(becauseopendataraisesquestionsabouthowthereleaseddataisselectedandimplicationsforthecity).59

59InthecaseofMontréal,someofthesequestionshavealreadybeendiscussedandexaminedindepthfollowingimplementationofMontréal’sOpenDataPolicyandDirectiveonDataGovernance.Thefactthatdatacollectedbythecityisownedbythecityservesasasolidfoundationforconsideringtheforegoingtopics.


February2018


Ratherthanpresentingtheresultsofasociologicalorpsychologicalstudy,MonfaredzadehandKrueger’sarticle(2015),InvestigatingSocialFactorsofSustainabilityinaSmartCity,underscorestheimportanceoftakingthesocialfactorintoaccountasaprerequisiteforasmartcityproject’ssuccess.

“Smartcities initiativesallowmembersofthecitytoparticipate inthegovernanceandmanagementofthecityandbecomeactiveusers.Anindividualmustbeabletoconnectinorder toachieveenhancementof socialandculturalcapitalaswellasachievemasseconomicgainsinproductivity.Iftheyarekeyplayerstheymayhavetheopportunitytoengagewiththeinitiativestotheextentthattheycaninfluencetheefforttobeasuccessorafailure”(MonfaredzadehandKrueger,2015,p.1113).

ThiskindofworkisfullyalignedwiththatofLewisMumford(1937)who,eveninhisday,arguedthaturbanplanningsufferedfromalackofsurveysonthecity’ssocialfunctions.ChrisLandford(2011)tackledthisproblembyproposingamatrixofvariablestoconsiderinassessingthesocialsustainability of an urban environment. Landford’s work does not, however, consider digitalinfrastructure.

Theacademicliteraturenowfavoursasharedcityapproachtobolsteringpublicparticipation.Thisschool is part of a global trend toward public participation, corresponding with ahypermodernistic version of public-spiritedness—a 21st century style of public engagementpromisinggainsinefficiencyandtransparency,whilenurturinghopesofsocialchange.

10.2.2 Use-PhaseStudies

Ifweslightlyexpandourfieldofresearch,weseethatengineeringuse-phasestudies(publishedsince the 1980s) also delve into the social acceptability of technologies and infrastructures,focusingontheuseoftheobjectunderstudy.Ause-phasestudyemploystimelinesshowinghowusersdo (ordonot)adoptaproduct, serviceor infrastructure (Terrade,etal.,2009).Theusephaseisoftenbrokendownintothreesequences:utility,usabilityandsocialacceptability.

“Utility”referstocorrespondencebetweenfunctionssupportedbythesystemanduser-assignedgoals. Put anotherway, utility is the partial or total conformity of the system’s featureswithcurrentorfutureusergoals.Usabilityistheeasewithwhichasystem’sfeaturescanbeusedandconsistsoffiveaspects:


February2018


1. Easeoflearning.2. Potentialperformance.3. Recollectionoffeatures.4. Errorprevention.5. Satisfaction.

In other words, “utility” is the correspondence between what the product, service orinfrastructureislikelytodoandwhattheuserwantsittodo,while“usability”referstoitseaseofuse(Tricot,etal.,2003).Sociology,scienceandmarketinghaveempiricallydemonstratedthataninnovationcanbeveryusefulandusablewithoutitbeingadoptedbyindividualsorsociety.

Studieshavealsodelved into thequestionofanobject’s socialacceptability,butwith far lessdetailorstructure.Suchresearchconsidersause’ssocioculturalcontext,whichcouldunderminethe system’s predicted acceptability, based on its utility and usability, in cases of certaininnovationsandundercertaincircumstances.

Whilemanyuse-phasestudiesexaminesocialacceptability,theydosofromtheperspectiveofindividualsandtheirinteractionsasauserwithaproductoraninfrastructure.Thisapproachisinconsistentwiththeprevailingdefinitionofsocialacceptability,whichemphasizescollective,notindividual,judgment.

Box11:AcceptabilityEvaluatedatThreePointsinTime

Usecanbeconsideredatthreepointsintime.● Before use: Evaluation before a person has used the product/infrastructure.

Acceptabilityreflectstheuser’ssubjectiveimpression,inconsiderationoftheproduct’sor infrastructure’s perceived: (1) utility, (2) perceived usability, (3) presumed socialinfluencesanddeploymentconditions.

● Afteruse:Evaluationinanexperimentalframework,onceanindividualhashadachancetousetheproduct/infrastructureatleastonce.Thesamethreefactorsareconsidered,withemphasisonthefirsttwo.

● Following integration into thedaily life: Evaluationof the service’sor infrastructure’sactualadoption,onceitismadeavailabletotheusersforinclusionintheirdailyroutines.The same three factorsare considered,withemphasison the first two (Tricot, et al.,2003).


February2018


10.2.3 FindingsonSocialAcceptabilityandtheUrbanIoTproject

TheliteraturereviewshowsthattheissueofsocialacceptabilityofurbanIoTprojects–whetherfrom a personal (in use-phase studies) or collective (in sociological studies) perspective–hasreceivedlessattentionthanethicalissues.

It is impossible at this stage to identify the smart city’s “social acceptability” issues withconfidence,eitheringeneralorwithrespecttoMontréal.However,thisreportlistsavarietyofissuesandconcerns thatcouldentailpublicopposition toanurban IoTprojectandconstituteimpedimentstotheproject’ssocialacceptability.

Sections 4 to 9 of the report describe ethical issues pertaining to privacy, social inclusion,separationofthegovernmentandbusinessspheres,transparency,reliabilityandfreedom,aswellas those concerning changes in governance modes that could trigger social opposition. Theliteraturealsoreferstootherconcerns,whichcannotbeclassifiedasethicalissues,becausetheyarenot clearlybasedonbasic socialprinciples (although thispoint isopen todiscussion),butcould be the subject of social opposition. Then there are concerns about certain IoT systemcomponents,inlightofrecenteventsinQuébec—suchasworriesabouttheproliferationofWi-Fitransmissions inpublicareas,opposition in2011 toHydro-Québec’s smartmeters,withmanycitizens speaking out against having more Wi-Fi waves passing through their homes—whichunderscoredsensitivitytothistopic.However,theliteraturereviewdidnotidentifyanyissuesofthiskindwithrespecttoIoT.

Figure 5, below, illustrates the ethical issues, alongwithmost of these concerns—or in otherwords, thevarious issueswith thepotential toharmtheproject’s socialacceptability. Section10.2.4,below,describestheidentifiedconcernsindetail.


February2018


Figure5:PotentialImpedimentstoSocialAcceptability

ThefollowingboxdescribestheonlypolltodateonhowQuebeckersfeelaboutIoT.Whilethetopiciscoveredgenerally(andnotspecifictoonecity),itdoeshighlighttrendsworthmentioning.

Box12:QuebeckersandIoT

AccordingtoCIRANO,QuebeckersbelievethatthebenefitsofusingtheInternetofThingsoutweightherisks.Seventy-threepercentofQuebeckerssupportusingthingsconnectedtotheInternetandsharingtheinformation,while45%believethatsuchuseissomewhatorhighlybeneficialtoQuébec.On the other hand, 20% of Quebeckers have little or no confidence in their government’smanagementofhowconnectedthingsareused.ThismeansabitlessthanonethirdofallQuebeckersfeelthereisariskinusingsuchtechnology.Between40and50%ofQuebeckerswouldbepreparedtosharedataontheirhealth,housing,travelanddrivingbehaviour(CIRANO,2017).

6

.

.

Enjeux'de'transforma0on

1.  Transforma+on,en,espace,prévisible,et,individualisé,

2.  Impact,sur,l’environnement,

Transforma+on,de,la,ville,

1.  Décisions,pour,l’op+misa+on,et,non,l’op+mum,social,

2.  Dilu+on,de,la,responsabilité,décisionnelle,

3.  Déterminisme,4.  Réduc+on,du,champ,des,

possibles,

Transforma+on,de,la,gouvernance,

Enjeux'éthiques

Inclusion,



Vie,privée,

Liberté,

Poten0els'freins'à'l’acceptabilité'sociale'


February2018


10.2.4 IdentifiedConcerns

Identified concerns that do not constitute ethical issues or pertain to changes in governancemodesaredescribedbelow.Suchconcernsareoftenthoseofthewritersandnotbasedonresultsofasociologicalorpsychosocialstudyofcurrentorfuturesmartcityresidents.

Table1:IdentifiedConcerns

Concerns Argument

Turningthecityintoa“predictablespace”

Many writers caution against over-planning urban trends, arguing that thisperspectiveischangingthecity’sroleasalived-inspaceanddiminishingqualityoflife.Algorithmsandtheprocessingofunderlyingdatareflectauniquevisionofthecity,drivenbythesesystemsthemselves.Technologicalarchitectureissimilartourbanabovegroundarchitecture,wherestreetsandbuildingsformspacesandvenues—oreliminatesuchspaces(Kitchin,2014).Recommendersystemsreducestimulation levels, exploration of new routes, chance encounters betweenresidents,etc.Theconstantsearchforefficiencythusappearsasanobstacletocitylife—inthesenseofundirectedactivityandwanderingaroundwithnospecificpurpose(Sassen,2011).60

Individualizationofthecity

Personalized access to certain servicesor information (such as advertising thattargets specific tastes, travel tips from downloaded apps based on recordedpreferences,etc.),givesrisetoinuniquepersonalurbanexperiences.Inthesameway,socialnetworkalgorithmsandsearchenginesgeneratespiralsofcontentthatcontinuously shrink tomore succinct and selective content.Widespreaduseoffilteringapps leadstodifferentexperiences inandadifferentunderstandingofthe city based on this vortexmodel. Developing “à la carte city”mechanisms,whereresidentsaredescribedasusersorclientsofservicesprovidedbythecity,givesthemaccesstoarangeoffreeandpaidservices(Baraud-Serfaty,2011).TheBordeaux3.0projectisoneexampleofthisphenomenon.

60OnepracticalexampleisanappthatpointsuserstoanyofVienna’spublictoilets.Whatistheriskthattouristsdiscoverfewersitesbystrollingaimlessly,becausetheapphasshownthemtheshortestpathtoarestroom?


February2018


Turningthepublicintoaconsumerofservices

Some writers also explain that the role of residents in smart city projects is,similarly,avectorofdepolitizationanddehumanization.Accordingtosmartcityrhetoric, the exercise of civic rights is often seen primarily from an economicperspective,withthecreationofservicesandapplicationssoldtopeoplewhoareperceivedonlyascustomers.Participationinurbansociopoliticallifeconsequentlytakes shape primarily in terms of creating products for sale, more than withrespecttocommunityaction.Residentsaccordinglyassumetheroleofconsumersofservicesofferedbycitygovernment,orbyotherindividualsorbusiness,ratherthan political players. This dynamic corresponds with an “uberization” of theeconomythatmanywritershavelambasted.Thiswidelypublicizedconceptrefersto the process of commercializing services and ties between members of thepublic(Morozov,2015).

Fearthatcitieswillbecomehomogeneousand“standardized”

Many writers have mentioned the homogenization that afflicts cities in theprocessofbecoming“smart.”Thisisbecausethestrengthsofcitieslielessintheir“IQscore”orinternationalclout,butintheirdistinctivefeatures,thewaystheystandoutandtheirareasofspecialization(Sassen,2011).Smartcityprojectsareoftensimilarlydesignedthroughouttheworld,becausetheyrelyontechnologicalresourcesthatarealikefromonecitytothenext(Poty,2014).Fromthisviewpoint,thesmartcityisacommercialproduct,withtheveryideadevelopedbyserviceproviders (Townsend, 2013). Consequently, there is a concern that “smart citysystems”suppliedbysuchproviderswillbereducedtogenericproductsthatfailtotreateachcityasaseparateentity,withitsowncharacteristics(sociocultural,economic,geographicandhistoricqualities)(Kitchin,2014).Itistheseveryfactors,however, which reveal each city’s real intelligence potential. This lack ofdifferentiation results from such factors as how the smart city project isunderstood. Categories developed to describe smart governments andinternationalclassificationsofacity’ssmartnessfailtoconsiderthesedistinctionsandproduceunequivocalvisions(Poty,2014).Nonetheless,whatworksinonecitymaynotinanother(Sassen,2011).

Thesmartcity’senvironmentalandenergycost

Smartinfrastructureoftenpermitsoptimizingtheuseofvariousresources,savingenergy and boosting efficiency. Digitalizing numerous services also makes itpossibletocutproductionandresourceconsumption(paperandstationery,officesupplies,etc.).Relocatingmunicipaljobscutscostsandconsumption(employeetravel,officeenergyconsumption,etc.)(Baraud-Serfaty,2015).Thesesavingsofenergyandresourcesareoftenemployedtotrumpetthemeritsofasustainablesmartcityproject.However,manywritershavecautionedagainstmereshiftsinconsumption.Suchshiftsoccurgeographicallywithinthecityanditssuburbswithrespecttotypesofconsumptionproducedbysmartcitytools.Thepropagationofservers, smart phones, computers and other connected objects, aswell as theresources deployed for their maintenance (coolers, uninterruptible powersupplies, generators, etc.), consume vast amounts of power (Viitanen andKingston,2013).Finally,thisshiftisoccurringworldwidewithanimpactonsocialandenvironmentaljustice.


February2018


11 Conclusion:TheSmartCity’sPromise

Thisreportprovidesanoverviewofacademicstudiesonissuesofethicalandsocialacceptabilityassociatedwithdeploymentofan InternetofThings inurban infrastructure.While studiesonethical issues pertaining to the Internet of Things are promising,we are at this pointmerelyconsideringresearcherhypothesesonthesmartcity’ssocialacceptability.Muchworkremains!

Wehavegivenequalweighttoalltheargumentscontainedinthisreport,sinceitwasimpossibletotakeintoaccountallofthethousandsofarticleswrittenonthistopic.Atthisstageofthework,wedonotwishtoweigharguments,solutionsorideas.Thatexerciseisessential,however,andwewilldosoinPhase2ofthisproject,workingincloseconjunctionwithMontréal’steam.Thisapproach is inspiredby our commitment tomaintain the impartiality expectedof a summaryreport.

Ourmission,asyouknow,isofthegreatestimportance.Wehaveachancetodeterminehowwellapromiseisreceived.

Apromiseisspeech.Whatisofferedbythepromiseiswords.Thevalueofsuchwordsareintrinsictothemselvesandnottothemeaningofthepromise.Inotherwords,itisthedeedthatcounts.Itisthedeedthatgivestime!Itgivestime,becausefromthedeed,theworldbecomestheWorldforthefirsttime.

Thepromisecreatesapotential,ratherthanafuture.Itfallswithintherangeofthepossible,butalso generates expectations. For Jacques Derrida, a promise has a three-part already-not-yetstructure.

“Of thepast, erased, singular event (anorigin, a trauma, anappointment), thereonlyremainsan inaccessible, lost trace,whichgives rise toa stillmasked truth, supposedlyknown, but concealed and not yet revealed. The promise, which is built around thisexpectation,supplementsandexceedsit.Wealwayspromisetoomuch.This‘toomuch’istheessenceofthepromise,makingitconfusinganddisturbing”(Derrida,2009).

JacquesDerridawrotethatthepromiseconstitutesexcess!An“excess”thatalwayscomplicatesthetaskof institutionsmaking it.Weshallreturntothistopic,aswewishtoexaminethe linkbetweenethicsandpromises,beforeproceedingfurther.

“Perhapsmorethaninothertypesofverbalperformanceutterance,thepromisecommitsitsmakertoafutureinvolvingitsidentityandresponsibilitywithrespecttothirdpartiesandsociety.Aswaryaspeopleareaboutpromises,theyremainacornerstoneofsociety.Theyareanappropriate—yetdisturbing—formofspeech.Thepromiseembodiestheforceneeded to drive all social life and is beyond the grasp of institutions” (Grieu, É. andThomasset,2005,p.75).


February2018


Thepromise,infact,isavehicleoftheethicsitnurtures:(i)perseverance,(ii)dutytoactand(iii)responsibility (Nachi,2003). Inadditiontothesecoreroles,AlainBoyeraddsthatthepromisecreatestheobligationofsomethingpromised,somethingowed.Indeed,thepromiseletsusseehintsinthepresentofnewlyemerginglife.

Aspreviouslynoted,all institutionsmakepromises.Weshallnowdiscusshowdifficult it is toacceptpromisesfrominstitutions,becauseoftheirsocialacceptability.Pastandcontemporarydisastershaveinoculatedusagainstabeliefinpromises.Thissentimentisnotonlyreflectedbyleadingschoolsofthought,butthesuspicionsaboutanyproposalthatwouldinfluencethecourseofevents.Suchpublicskepticismofpromisesismagnifiedbyourawarenessofthefragilityoflife.Thosewhobelievesweepingpromisesriskbeingperceivedasnaiveortruebelievers.

Makingapromisemeansaskingsomeonetotrustthatsomethingwillcometrueinthefuture.Itisonlyovertimethatwecandetermineifthepromisewaskept.Apromiseisuniqueinthatthemereideaofanexcuseisinapplicable.ThatiswhywemustnowcarefullyconsiderthedifferentoptionswehavedescribedtoaddresstheethicalissuesandsocialacceptabilityoftheInternetofThingsinprovidingourbestsupportinmakingtheVilledeMontréalaniconofthesmartcity’spromise.


February2018


12 Bibliography

Abiteboul,S.andFroidevaux,C.(2016),“Autourdel’informatique:algorithmesetladisparitiondusujet.Interview with A. Rouvroy,” The Conversation. Viewed at: http://theconversation.com/autour-de-linformatique-les-algorithms-et-la-disparition-du-sujet-53515.ViewedFebruary20,2018. Amossy,R. (1989),“Lanotiondestereotype intheréflexioncontemporaine,”Mutationsd'images.73:pp.29-46. Ananny,M.andK.Crawford(2016),“Seeingwithoutknowing:Limitationsofthetransparencyidealanditsapplicationtoalgorithmicaccountability,”NewMedia&Society:https://doi.org/10.1177/1461444816676645Angelidou,M.(2014),“Smartcitypolicies:Aspatialapproach,”Cities41,Supplement1:S3-S11.Baraud-Serfaty,I.(2011),“Lanouvelleprivatisationdesvilles,”Esprit3(March/April):149-167. Barocas,S.andH.Nissenbaum(2014),“BigData'sEndRunaroundAnonymityandConsent,”Privacy,BigData,andthePublicGood:FrameworksforEngagement.J.S.Lane,Victoria.Bender,Stefan.Nissenbaum,Helen,eds.NewYork,NY,CambridgeUniversityPress:pp.45-75. Batini, C., Cappiello, C., Francalanci, C., and Maurino, A. (2009), “Methodologies for data qualityassessmentandimprovement,”ACMComputingSurveys(CSUR)41(3). Battelier,P.(2015),“Acceptabilitésociale,cartographied'unenotionetdesesusages,”Cahierderesearch,UQAM:LespublicationsduCentr'ERE.Bensinger,G.(2014),“AmazonWantstoShipYourPackageBeforeYouBuyIt,”TheWallStreetJournal.Jan.17,2014. Berthier, T. andO.Kempf (2016), “Versunegéopolitiquede ladonnée,”Réalités industrielles.August2016. Birman,J.(2011),“Jesuisvu,doncjesuis:lavisibilitéenquestion,”Lestyranniesdelavisibilité.Aubert,N.andHaroche,C.,eds.Toulouse,Èrescoll,”Sociologieclinique,pp.39-52.Brender, N. (2012), “Étude du dilemme urbain : urbanisation, pauvreté et violence,” C. Summarydocument.WrittenbyNatalieBrender,basedonthestudybyRobertMuggah,CRDI. Breux,S.andJ.Diaz(2017),“Lavilleintelligente:origine,définitions,forcesetlimitesd'uneexpressionpolysémique,”U.U.d.recherché,Montréal,Institutnationaldelarecherchescientifique.


February2018


Brunson, M., et al. (1996), Defining social acceptability in ecosystem management: a workshopproceedings.Gen.Tech.Rep.PNW-GTR-369.,Portland,OR. Bunker,E.(2016),Aucunebêteaussiféroce,PayotandRivages. Carcassone,G.(2001),“Letroubledelatransparence,”Pouvoirs97:17-23. Cardon,D.(2015),Aquoirêventalgorithms-Nosviesàl'heuredesbigdata.Paris,LeSeuil. Cate,Fred.H.(2006),“TheFailureofFairInformationPracticePrinciples,”ConsumerProtectionintheAgeoftheInformationEconomy,pp.343-379.Canguilhem,G. (2015) “Lenormalet lepathologique,”Paris,PressesUniversitairesdeFrance,Quadrige,2015,12thedition.Champeau,G. (2016), “Détecter les futurs terroristes sur Internet ? L’Europe veut s’inspirer d’Israël.”Elnet. CIRANO(2017),Baromètredesperceptionsdesquébécois,PresseinternationalePolytechnique.Availableat:https://barometre.cirano.qc.ca/pdf/issues_sante.pdf Citron,D.(2010),“TechnologicalDueProcess.”WashingtonUniversityLawReview85:1249-1256. Clarke,R.(1998),“InformationTechnologyandDataveillance,”CommunicationsoftheACM31(5):198-512. Collins,H.(2003),EnterpriseKnowledgePortals,AMACOM.Commissiondel’éthiqueenscienceettechnologiesQuébec(2017),“L’éthiqueetlamorale,dequoionparle?”RetrievedJuly22,2017,fromhttp://www.ethique.gouv.qc.ca/fr/ethique/quest-ce-que-lethique/lethique-et-la-morale-de-quoi-on-parle.html.Commissiondelascienceandlatechniqueauservicedudevelopment.(2016),Infrastructuresetvillesintelligentes,UNEconomicandSocialCouncil,UnitedNations. CrawfordK.andSchultz,J.(2014),“Bigdataanddueprocess:towardaframeworktoredresspredictiveprivacyharms,”BostonCollegeLawReview55(93):93-130.D'Elbée, P. (2015), “L'exigencede l'humain face à la technique,”Raisonnance. Cahier de réflexiondesmairesfrancophones(6):22-25. Delain,P.(2017),LesmotsdeJacquesDerrida,Ed:Guilgal,2004-2017.PagecreatedAugust25,2005.


February2018


Delain,P.(2017),Lapromesse.1.Archi-promesse,Ed.:Guilgal,2004-2017,PagecreatedJuly5,2009. Deleury,E.(2016),“Villeintelligente:lenumériqueetl'éthiquedoiventallerdepair,”HuffingtonPost,2016-02-24. Desmond,M.(2012),“EvictionandtheReproductionofUrbanPoverty,”AJS118(1):pp.88–133. Doran,M.-A.(2014),“Démystifierlesvillesetcommunautésintelligentes,”LeSablier21(1):pp.20-29. Douay,N.andC.Henriot(2016),“LaChineàl’heuredesvillesintelligentes,”L'Informationgéographique3(80):pp.89-102. Dusek,V.(2006),PhilosophyofTechnology:AnIntroduction.Oxford,BlackwellPublishing. Erlingsson,U.e.P.,VasylandKorolova,Aleksandra(2014),“RAPPOR:RandomizedAggregatablePrivacy-Preserving Ordinal Response,” Proceeding of the 2014 ACM SIGSAC Conference on Computer andCommunicationsSecurity.Scottsdale,Arizona. FederalTradeCommission(2015),Internetofthings:Privacyandsecurityinaconnectedworld,p.71.Felli,R.(2015),“Ladurabilitéoul’escamotagedudéveloppementdurable,”Raisonspolitiques4(60):149-160. Fijalkow,Y.(2007),“Sociologiedesvilles,”Repèressociologie.C.LaDécouverte. Floridi,L.(2012),“Bigdataandtheirepistemologicalchallenge,”Philosophy&Technology25(4):435-437. Foucault,M.(1975),Surveilleretpunir.Naissancedelaprison,Paris,Gallimard. Gendron, C. (2014), “Penser l’acceptabilité sociale : au-delà de l’intérêt, les valeurs,” Communicate,pp.117-129. Germain,A.(1997),“L’étrangeretlaville,”CanadianJournalofRegionalScienceSpring,summer:pp.237-254. Goffman,A.(2009),“OntheRun:WantedMeninaPhiladelphiaGhetto,”AmericanSociologicalReview74(3):pp.339-357. Goodman, B., Flaxman S. (2016), “EU regulations on algorithmic decision-making and a ‘right toexplanation,’ presented at 2016 ICMLWorkshoponHuman Interpretability inMachine Learning (WHI2016),NewYork,NY. Goodman,B.(2016),“AStepTowardsAccountableAlgorithms?:AlgorithmicbiasandtheEuropeanUnion


February2018


General Data Protection,” 29th Conference on Neural Information Processing Systems (NIPS 2016),Barcelona,Spain. Greenfield, A. (2014), “The truth about smart cities: ‘In the end, they will destroy democracy,’” TheGuardian. Viewed at: https://www.theguardian.com/cities/2014/dec/17/truth-smart-city-destroy-democracy-urban-thinkers-buzzphrase Grieu, É. T., A. (2005), “La promesse à la source des relations interpersonnelles et sociales,” Revued'éthiqueandthéologiemorale236:pp.55-76. HarvardLawSchool,etal.(2017),TheFutureofIoT:Summaryreport.HarvardLawSchool,inpartnershipwiththeKnightFoundation. Haggerty,K.D.andEricson,R.V.(2006),TheNewPoliticsofSurveillanceandVisibility.Toronto,UniversityofTorontoPress. Heisenberg,W.(1990),LaPartieetleTout-LeMondedelaphysiqueatomique.Paris,Flammarion. Herschel,R.,Miori,V.(2017),“EthicsandBigData,”TechnologyinSociety49:31-36. Inspire (2015), EU INSPIRE Directive for Spatial Data. Available at:https://inspire.ec.europa.eu/Legislation/Spatial-Data-Services/580 InstitutdelastatistiqueduQuébec(2015),“Lescompétencesenlittératie,ennumératieetenresolutiondeproblemsdanslesenvironnementstechnologiques:desclefspourreleverlesdéfisdu XXIe siècle,” Rapport québécois du Programme pour l’évaluation internationale descompétencesdesadultes(PEICA),Québec,QC,InstitutdelastatistiqueduQuébec:p.247.Metcalf, J. (2014), Ethic Codes: History, Context, and Challenges, draft, viewed at:https://bdes.datasociety.net/council-output/ethics-codes-history-context-and-challenges/. ViewedFebruary20,2018 Kaplan,D.(2012),Taville,tropsmartpourtoi,viewedathttp://www.internetactu.net/2012/10/02/ta-ville-trop-smart-pour-toi/.ViewedFebruary20,2018. Kerr, I.andJ.Earle (2013),“Prediction,Preemption,Presumption:HowBigDataThreatensBigPictureprivacy,”StanfordLawReviewOnline66(65):pp.65-72. Kitchin,R.(2014),“Therealtimecity.Bigdataandsmarturbanism,”GeoJournal79(1):pp.1-14. Kitchin,R.(2015),“Makingsenseofsmartcities:Addressingpresentshortcomings,”CambridgeJournalofRegions,EconomyandSociety8(1):pp.131-136.


February2018


Kitchin,R.(2016),“Theethicsofsmartcitiesandurbanscience,”PhilosophicalTransactionoftheRoyalSociety374(2083). Kitchin,R.(2016),“Thinkingcriticallyaboutandresearchingalgorithms,”Information,Communication&Society:pp.1-16. Kitchin,R.a.(2014),Thedatarevolution:bigdata,opendata,datainfrastructures&theirconsequences,London:SagePublicationsLtd,2014,©2014. Koolhaas, R. (2014), “My thoughts on the smart city,” Digital Minds for a New Europe, EuropeanCommission. Sweeney,L.(2013),“Discriminationinonlineaddelivery,”ACMQueue11(3). Landorf,C.(2011),“Evaluatingsocialsustainabilityinhistoricurbanenvironments,”InternationalJournalofHeritageStudies,17(5):pp.463-477. Larson, J. e. M., Surya, Kirchner, Lauren and Angwin, Julia (2016), How We Analyzed the COMPASRecidivism Algorithm. Viewed at: https://www.propublica.org/article/how-we-analyzed-the-compas-recidivism-algorithm.ViewedMay23,2016 Le Galès, P. (1995), “Gouvernement des villes à la gouvernance urbaine,”Revue française de sciencepolitique45(1):pp.57-95. Lecomte,S.(2017),“Ledroitàl’oublinumérique:regardscroiséssurlalégislationapplicableenEurope,auCanadaetauxÉtats-Unis,”CanLIIConnecte. Mannermaa,M. (2007), “Living in theEuropeanUbiquitousSociety,” Journalof Futures Studies 11(4):pp.105-120.Mantelero, A. (2014), “The future of consumer data protection in the EU: Rethinking the ‘notice andconsent’ paradigm in the new era of predictive analytics,” Computer Law and Security Report 30(6):pp.643-660.Gaughan,M.(2016),PrivacyintheSmartCity:Implicationsofsensornetworkdesign,law,andpolicyforlocationalprivacy,Master’sthesis.UrbanStudies,UniversityofWashington.Metcalf,J.andK.Crawford(2016),“WherearehumansubjectsinBigDataresearch?Theemergingethicsdivide,”BigData&SocietyJanuary-June:pp.1-14.Michaud,T.(2010),“Lascience-fiction :uneculturedeinnovationglobale,”JournalforCommunicationStudies3(1):pp.171-180.


February2018


Mittelstadt,B.D.e.A.,PatrickandTaddeo,Mariarosaria,Wachter,SandraandFloridi,Luciano(July-December 2016), “The ethics of algorithms: Mapping the debate,” Big Data & Society,https://doi.org/10.1177/2053951716679679Mohanty, S. P., et al. (2016), “Everything you wanted to know about smart cities,” IEEE ConsumerElectronicsMagazineJuly:pp.60-70.Monfaredzadeh,T.,RobertKrueger,(2015), Investigatingsocialfactorsofsustainabilityinasmartcity.InternationalConferenceonSustainableDesign,EngineeringandConstruction,ProcediaEngineeringMoreno, C. (2015), “Ville intelligente : citoyenne et connectée,”Raisonnance. Cahier de réflexion desmairesfrancophones.(6):pp.8-12.Morozov, E. (2014), “De l’utopie numérique au choc social,” Le Monde Diplomatique. Viewed at:https://www.monde-diplomatique.fr/2014/08/MOROZOV/50714.ViewedFebruary20,2018.Morozov, E. (2015), “Résister à l’uberisation du monde,” Le Monde Diplomatique. Viewed at:https://www.monde-diplomatique.fr/2015/09/MOROZOV/53676.ViewedFebruary20,2018. Morozov,E.(2015),“Lemiragenumérique.PourunepolitiqueduBigData,”Paris,LesPrairiesordinaires. Morozov, E. (2016), “La Sécu selon Uber,” Le Monde Diplomatique. Viewed at:https://blog.mondediplo.net/2016-09-16-La-Secu-selon-Uber.ViewedFebruary20,2018. Moscovici,S.(2014),PsychologieSociale,PressesUniversitairesdeFrance. Mossenburg,K.,Tolbert,J.andStansbury(2003),VirtualInequality:BeyondtheDigitalDivide.WashingtonDC,WashingtonUniversityPress. Mumford,L.((1937)2000),“WhatisaCity?”Thecityreader.R.T.e.F.S.LeGates.London,Routledge. Nachi,M.(2003),Éthiquedelapromesse:L’Agirresponsable.Paris,PressesUniversitairesFrance. NarayananA.,H.J.,FeltenE.W.(2016),“APrecautionaryApproachtoBigDataPrivacy,”DataProtectionontheMove.Law,GovernanceandTechnologySeries,L.R.GutwirthS.,DeHertP.Dordrecht,Springer.p.24. Conseilnationaldunumérique(2013),“Citoyensd'unesociéténumérique.Accès, littératie,médiation,pouvoird'agir:pourunenouvellepolitiquedel'inclusion,”RapportàlaMinistredéléguéechargéedespetitesetmoyennesentreprises,del'Innovationandl'Économienumérique.Paris:p.88.


February2018


Oddoux,A.(2016),“«Smartcity»:dequelleintelligenceparle-t-on?”BloguedeMarcoCremaschi:Cycleurbanisme2016-2017:Noschroniques.Viewedat:https://cremaschiblog.wordpress.com/2016/11/12/smart-city-de-quelle-intelligence-parle-ton-anais-oddoux/.ViewedFebruary20,2018. ODI (2015),OpenDataCertificate, viewedat: https://certificates.theodi.org/en/.ViewedFebruary20,2018. Ohm,P. (2010),“BrokenPromisesofPrivacy:Responding to theSurprisingFailureofAnonymization,”UCLALRev:1701-1777. EuropeanParliament(2015),“Bigdataandsmartdevicesandtheirimpactonprivacy,”StudyfortheLIBECommittee.C.s.r.a.c.affairs,EuropeanParliament. EuropeanParliament(2016),Regulation(EU)2016/679oftheEuropeanParliamentandoftheCouncilof27April2016ontheProtectionofNaturalPersonswithRegardtotheProcessingofPersonalDataandtheFreeMovementofSuchData,O.J.o.t.E.Union. Pedreschi,D.e.R.,S.andTurini,F.(2009),“Integratinginductionanddeductionforfindingevidenceofdiscrimination,”InternationalConferenceonArtificialIntelligenceandLawACM:pp.157-166. Peres,É.(2015),“Lesdonnéesnumériques:unenjeud’educationandcitoyenneté,”JournalofficieldelaRépubliquefrançaise.AvisduConseiléconomique,socialetenvironnemental,p.154. Picon,A.(2013),Smartcities.Théorieetcritiqued’unidéalauto-réalisateur,Paris,éditionsB2. Pisani,F.(2015),“Maisd'oùvientcetteidéebizarrede«villeintelligente»?LeblogLaTribune,2017.Viewedat:https://www.latribune.fr/blogs/aux-coeurs-de-l-innovation/20150116trib4e9bdc2e1/mais-d-ou-vient-cette-idee-bizarre-de-ville-intelligente.html.ViewedFebruary20,2018. Plantard,P.(2013),“E-inclusion:Braconnage,bricolageetbutinage,”Placepublique(Dossier:Faut-ilavoirpeurdelavillenumérique?),pp.17-21. Polonetsky,O.T.a.J.(2013),“BigDataforAll:PrivacyandUserControlintheAgeofAnalytics,”Nw.J.Tech.&Intell.Prop11(239). Poty,P.(2014),“Smartcities.Lesclésnumériquespourlavilleintelligente,”PlateformepourlaWallonienumérique. Viewed at: https://www.digitalwallonia.be/smartcities-cles-numeriques-pour-la-ville-intelligente/.ViewedFebruary20,2018. McArdle, R and R. Kitchin (2014), “Improving the Veracity of Open and Real-Time Urban Data,” TheProgrammableWorkingCityPaper.p.13.


February2018


Raley,R.(2013),“DataveillanceandCountervailance,”RawDataIsanOxymoron,L.Gitelman,Cambridge,MITPress:pp.121-145. Rallet,A.andRochelandet,F. (2004),“Lafracturenumérique :unefaillesans fondement?”Networks5(127-128):pp.19-54. Richards,N.M.(2013),“TheDangersofSurveillance,”HarvardLawReview126(1934):1934-1965. Richards,N.M.andJ.H.King(2014),“BigDataEthics,”WakeForestLawReview49:pp.393-432. Ricoeur,P.(2000),LaMémoire,l’histoire,l’oubli,Paris,Seuil. Romei, A. and Ruggieri, S. (2013), “Discrimination Data Analysis: A Multi-disciplinary Bibliography,”DiscriminationandPrivacy in the Information Society. B. Custers, Calders, T., Schermer, B., Zarsky, T.;Heidelberg,Springer-VerlagBerlin. Rouvroy,A.(2008),“Réinventerl’artd’oublierandsefaireoublierinthesociétéinformation?”Lasecuritydel’individunumérisé.Réflexionsprospectivesetinternationales.S.Lacour.Paris,L’Harmattan:249-278. Rouvroy,A.e.Stiegler,B.(2016),“TheDigitalRegimeofTruth:FromtheAlgorithmicGovernmentalitytoaNewRuleofLaw,”LADELEUZIANA–ONLINEJOURNALOFPHILOSOPHY. Rubinstein,I.S.(2013),“Bigdata:theendofprivacyoranewbeginning?”InternationalDataPrivacyLaw3(2):pp.74-87. RUMPALA,Y.(2010),“Cequelascience-fictionpourraitapporteràlapenséepolitique,”Raisonspolitiques4(40):p.93-112. Sandovig,C.andK.Hamilton,K.Karahalios,C.Langbort(2014),“AuditingAlgorithms:ResearchMethodsforDetectingDiscriminationonInternetPlatforms,”DataandDiscrimination:ConvertingCriticalConcernsinto Productive Inquiry,” at the Preconference of the 65th Annual Conference of the InternationalCommunicationAssociation.Seattle. Sassen,S.(2010),Talkingbacktoyourintelligentcity.McKinseyonSociety,eds. Schelings,C.,Elsen,C.(2017),SmartCityconcepts:fromperceptiontoacceptance,21stConferenceoftheEnvironmentalandSustainabilityManagementAccountingNetwork(EMAN),Liège,Brussels. Seth, S. (April 5, 2017), In How Many Ways Can an Algorithm be Fair? Viewed at:https://www.turing.ac.uk/events/many-ways-can-algorithm-fair-talk-visiting-researcher-suchana-seth/.ViewedFebruary20,2018


February2018


Smolar,P.(2016),“AlgorithmessécretsdelasurveillanceisraélienneenCisjordanie,”LeMonde.Viewedat: http://www.lemonde.fr/international/article/2016/11/25/les-algorithms-secrets-de-la-surveillance-israelienne-en-cisjordanie_5037999_3210.html.ViewedFebruary20,2018. Söderström, O., et al. (2014), “Smart cities as corporate storytelling,” City. Analysis of urban trends,culture,theory,policy,action18(3):pp.307-320. Solove,D.(2004),TheDigitalPerson:TechnologyandPrivacyintheInformationAge.NewYork/London,NewYorkUniversityPress. Stahl,B.C. (2011),“IT forabetter future:howto integrateethics,politicsand innovation,” JournalofInformation,Communication&EthicsinSociety9(3):pp.140-156. Stoycheff,E.(2016),“UnderSurveillance:ExaminingFacebook’sSpiralofSilenceEffectsintheWakeofNSAInternetMonitoring,”Journalism&MassCommunicationQuarterly93(2):pp.1-16. EDPS–EuropeanDataProtectionSupervisor(2014),Privacyandcompetitivenessintheageofbigdata:theinterplaybetweendataprotection,competitionlawandconsumerprotectionintheDigitalEconomy.Brussels:p.41. Tavani, H. (2004), Ethics and Technology: Ethical issues in an age of information and communicationtechnology.Hoboken,N.J.,JohnWilleyandSons. Taylor,A.(2014),Démocratie.com.Montréal,Lux. Terrade, F., et al. (2009), “Social acceptability: How social determinants can influence analysis oftechnologysystemacceptability,”TravailHumain72(4),pp.383-395. IERC—European Research Cluster on the Internet of Things (2015), Internet of Things: IoTGovernance, Privacy, and Security Issues, E. C. I. S. a.Media. Brussels, EuropeanCommission:InformationSocietyandMedia. Thomas,L.-V.(1984),Fantasmesauquotidien,Paris,Méridiens. UNHCHR(2014),“Therighttoprivacyinthedigitalage:ReportoftheOfficeoftheUnitedNationsHighCommissioner forHumanRights,”Annual reportof theUNHCHRand reportsof theOfficeof theHighCommissionerandtheSecretary-General.UNHCHR.Geneva,UNHCHR. US EPA (2006), Data Quality Assessment: A Reviewer’s Guide, US EPA, Office of EnvironmentalInformation,WashingtonDC,February2006. VandenHoven,J.,etal.(2012),“FutureICT-theroadtowardsethicalICT,”TheEuropeanPhysicalJournalSpecialTopics214:pp.153-181.


February2018


VandenHoven,J.(2016),Factsheet-EthicsSubgroupIoT-Version4.0.E.Union.Ville de Montréal (2017), Montréal, smart city et numérique : Stratégie montréalaise 2014-2017.Montréal,VilledeMontréal.Walker,N.R.(2016),“AmericanCrossroads:GeneralMotors’MidcenturyCampaigntoPromoteModernistUrbanDesigninHometownU.S.A.Buildings&Landscapes,”JournaloftheVernacularArchitectureForum,Volume23,Number2,Fall2016,pp.89-115.Wilson,W.J.(1987),TheTrulyDisadvantaged:TheInnerCity,theUnderclass,andPublicPolicy,Chicago,UniversityofChicagoPress. Schwab,K.(2017),TheFourthIndustrialRevolution:whatitmeans,howtorespond.WorldEconomicForum.Availableat:https://www-weforum-org.proxy.bibliotheques.uqam.ca:2443/agenda/2016/01/the-fourth-industrial-revolution-what-it-means-and-how-to-respond/Ziegeldorf,J.H.O.G.M.K.W.(2014),“PrivacyintheInternetofThings:ThreatsandChallenges.”Securityandcommunicationnetworks7:pp.2728-2742. Zook,M.,Barocas,S.,Boyd,D.,Crawford,K.,Keller,E.,GangadharanS.P.,etal.,e1005399(2017),“Tensimplerulesforresponsiblebigdataresearch,”Editorial,PLOSComputationalBiology13(3).


February2018


AppendixASupplementtoSectiononPrivacy

IsProtectingPrivacyImportant?

Privacyisfundamentaltotheproperfunctioningofasociety.Bylettingpeoplekeepinformationconfidential, a society permits independent action, personal thinking and experimentation.Privacy nurtures diversity, development and the ability to copewith social pressures (Solove,2006).

HowDoesthePublicFeelAboutIoT-PrivacyConcerns?

The2015TRUSTeUSConsumerPrivacyConfidenceIndex, indicatedthat20%ofonlineserviceusersbelievethebenefitsassociatedwiththe InternetofThingsaremore important thantheprivacyconcernstheyengender(TRUSTe,2015).WhilethesefindingspertaintoEuropeandtheInternetofThingsinvolvingcommonandpersonalobjects(watches,medicalequipment),theyimplyaveryrealanxietywiththegeneraluseoftechnology.

WhatIsPersonalData,inTermsoftheLaw?

Theconceptofprivacyisrelatedtothatofpersonaldata.Canadianlegislationdefinesthelatteras information about an identifiable individual, and in particular, information pertaining to aperson’sidentity(race,religion,education,background,identifyingdata,address,aswellasothercriteria),his/heropinionsorpersonalideasandtheideasandopinionsofotherabouthim/her.Formoreinformation,visit:http://laws-lois.justice.gc.ca/fra/lois/P-21/.

WhatistheTheoryofContextualIntegrity?

HelenNissenbaumproposedthatprivacyisprimarilybasedonthecontextinwhichinformationis exchanged, which gives rise to rules for such exchanges (Nissenbaum, 2004; Barocas andNissenbaum, 2014). Informationmay ormay not be sharedwithout infringing on someone’sprivacy, depending on an interplay of multiple factors, including the parties’ relationship,sensitivityoftheinformationanddirectionoftheexchange(two-wayorone-way),61asillustratedinthefollowingfigure.Privacyisnotaneither/orsituation,asitdependsoncontext,ratherthanthe kinds of information communicated. Nissenbaum (2014) consequently suggested thatindividualsmightbeentitledtoprivacyinapublicspace.

61Thisiscalled“theprincipleofcontextualintegrity”(BarocasandNissenbaum,2014).Forexample,rulesonhow information isused inahealthcareestablishmentdeterminewhat kindsof information canbesharedby thepartiesconcerned (patient,doctor,administrativestaff, family). In this situation,patientswho provide access to their personal information can do so, while maintaining their privacy, if theinformation is handled according to rules and social expectations pertaining to disclosure, sharing andconfidentiality.


February2018


Figure6:ContextualPrivacyFactors(Gaughan,2016,p.17)

AreThereAnyCasesofAnonymizedDataBeingRe-Identified?

Acquisti,GrossandStutzman(2011)demonstratedthatitisnowpossibletoascertainaperson’ssocialinsurancenumberusingaphotoofhisorherface.SeveralstudiesbySweeney(2000;2013)alsorevealthepossibilityofre-identifyingindividualswithpublicdata.Meta-andgeolocationdata,itshouldbenoted,playkeyrolesinidentifyingindividuals(Montjoye,etal.,2013).

Therehavebeenseveralcaseswherere-identifieddatathatwasreleasedpubliclywasabletobere-identified,orwheredatathatwasassumedtohavenoidentifyingfeaturescouldbecorrelatedwithspecificpopulations.Forexample,in2013,theNewYorkCityTaxiandLimousineCommissionreleasedadatasetof173million individual cab rides,and it included thepick-upanddrop-offtimes, locations, fare and tip amounts. The taxi drivers’medallion numberswere anonymized(hashed)butthiswasquicklyde-anonymized–revealingsensitiveinformationsuchasanydriver’sannual income and enabling researchers to infer their home address (Franceschi-Bicchierrai,2015).AdatascientistatNeustarResearchshowedthatbycombiningthisdatasetwithotherformsofpublic information like celebrityblogsyoucould trackwell-knownactors,andpredictlikely home addresses of people who frequented strip clubs (Tockar, 2014, in Metcalfe andCrawford,2016).AnotherresearcherdemonstratedhowthetaxidatasetweredevoutMuslimsbyobservingwhichdriversstoppedatMuslimprayertimes(Franceschi-Bicchierrai,2015).

WhatAretheBasicPrinciplesofPrivacybyDesign?

PrivacybyDesigninvolves:1)Proactive(ratherthanreactive)solutions.2)Privacyprotectionbydefault.3)Privacyprotectionintegratedintothedesign.4)Fullfunctionality,sosystemscancollecthighqualitydata.5)End-to-endsecurity,throughouttheentirelifecycle.6)Visibilityandtransparency.7)Respectforuserprivacy(user-centredsystemdesign)(Gaughan,2016,p.57).62

62Note to self—SRG:PrivacybyDesign -14 (AnnCavoukian,PrivacybyDesign.Fromrhetoric to reality.2014).


February2018


StrategiesProposedforTranscendingtheLimitationsofAnonymization

The literature provides several solutions for transcending the limitations of anonymization,including:

● Organizationalcommitmentnottore-identifydata.● Aggregate(coarse-grained)datacollection.● Minimizeddatacollection.

Inits2015recommendations,theFederalTradeCommissionsuggestedthatcompanieskeeptheirdata in de-identified form (FTC, 2015) and implementmechanisms to ensure long-term datade-identification.Accordingtothisapproach,businessesshould:1)Takereasonablemeasurestode-identify data. 2)Make a public commitment not to re-identify data. 3) Enter into bindingcontractswiththirdpartiesthatbandatasharingandincludethecommitmentnottore-identifydata (FTC, 2015). As Tene andPolonetsky suggested (2013), the FTChasmoved away fromastrategyfocusedonthedegreeofdataidentifiabilitytoonebuiltaroundorganization’sintentionsanditscommitmenttopreventre-identification.Thisapproachtakesintoaccountthatfactthatdataisnotsimplyprivateornon-private,butfallswithinaconstantlyredefinedprivacycontinuum(Tene and Polonetsky, 2013). Means of curbing and monitoring possible redefinition mustaccordinglybefound.

Anotherproposedalternative is collectingcoarse-graineddata (vandenHoven,2012,p.170).Gaughan(2016)gaveanexampleofdatacollectionsystemsthatcanbeconfiguredaspartofa“local aggregation technology” to aggregatedata at the sourceandeliminate storageof localdisaggregated data that could be associated with individuals (Gaughan, 2016, p. 60). Thisapproachhelps reducequantities of stored sensor networkdata,with aggregating algorithmstransforming individual data into summary statistics, based on system designer settings(Gaughan, 2016, p. 60; Narayanan, 2016). Under such circumstances, no individual could begeolocatedandinmostcases,nodalaggregationstatisticscouldmeettheanalyticalneedsofcitygovernments.

Minimizeddatacollectionisalsooftenrecommended(Narayanan,2016).

WhatNewApproachesAreBeingConsideredforProtectingPrivacywithMetadata?

ResearchersfundedbytheEuropeanCommissionarenowworkingonimplementingstickyflowpolicies that would enforce access and confidentiality policies using metadata to mark datastreams.Thiscouldmean,forexample,thatdatabemarkedwithasecuritypolicydescribinghowitcanbeusedandwhatconditionsmustbemetbeforeitcanmigratetoanewdataprocessingunit(IERC,2015,p.52).


February2018


WhatAreCitron’sAdditionalRecommendations(2010)onAutomatedDecision-Making?

DanielleCitron(2010)proposedavarietyofadditionalmechanismsthatcouldbeapplied,suchas:

● Investing in training on unconscious bias and automation bias for employees usingsystems tomake administrative decisions. Such trainingwill help thembecomemorecriticalofdecisionsmadebysuchsystems.

● Requireuserentitiestoprovidedetailedexplanationsofdecisionsmadebyautomatedsystems,includingcomputer-generatedinformationandresults.

● Requireuserentitiestotestsystemsoftwareregularlyforbiasandothererrors(Citron,2010).

CouncilforBigData,Ethics,andSociety’s10RulesforResponsibleBigDataResearch

Tenrulesforresponsiblebigdataresearch(Zook,etal.,2017)

ModeledonPLOSComputationalBiology–thefirstfiverulesaroundhowtoreducethechanceofharmresultingfrombigdataresearchpractices.Thesecondfivefocusonwaysresearcherscancontribute tobuildingbest practices that fit their disciplinaryandmethodological approaches.Paper is fromtheCouncil forBigData,Ethics,andSociety,agroupof20scholars fromawiderangeofsocial,naturalandcomputationalsciences.

1. Acknowledgethatdataarepeopleandcandoharm:datarepresentandimpactpeople2. Recognizethatprivacyismorethanabinaryvalue:privacyiscontextual[11]and

situational[12],notreducibletoapublic/privatebinary.Privacyalsogoesbeyondsingleindividualsandextendstogroups[10].Thisisparticularlyresonantforcommunitieswhohavebeenonthereceivingendofdiscriminatorydata-drivenpolicieshistorically,suchasthepracticeofredlining[14,15,16]

3. Guardagainstthere-identificationofyourdata[22](Evendataaboutgroups(aggregatestatistics)canhaveseriousimplicationiftheyrevealthatcertaincommunities,forexample,sufferfromstigmatiseddiseasesorsocialbehaviourmuchmorethanothers[27]Identifypossiblevectorsofre-identificationinyourdata(MOREHERE)

4. Practiceethicaldatasharing5. Considerthestrengthsandlimitationsofyourdata.bigdoesnotautomaticallymean

better6. Debatethetough,ethicalchoices:“ratherthanabug,thelackofclear-cutsolutionsand

governanceprotocolsshouldbemoreappropriatelyunderstoodasafeaturethatresearchersshouldembracewithintheirownwork”(Zook,etal.,2017,5)

7. Developacodeofconductforyourorganisation,researchcommunityorindustry:asameanstocementthisindailypractice.

8. Designyourdataandsystemsforauditability9. Engagewiththebroaderconsequencesofdataandanalysispractices:10. Knowwhentobreaktheserules:intimesofemergency

(Zook,etal.,2017)


February2018


However,thefieldofresearchethicsisgrapplingwiththeinabilityofexistingframeworkstodealwith issues raisedbybig data and,more generally, Economy4.0. Researchonbig data is notconcernedwithphysicalharmtopeople,buttoissuesofprivacyanddiscriminationresultingfromdataanalysis.Thisisbeyondtheusualscopeofresearchethics,whichhasbeengreatlyinfluencedby the biomedical field. Such a framework also often focuses on how choices aremade andconsentisgivenduringdatacollection.However,suchchoiceandconsentruleshavenotbeeneffectivelyappliedand,ontheirown,arenolongerrelevant.Furthermore,researchwithbigdata“Itfundamentallychangesourunderstandingofresearchdatatobe(atleastintheory)infinitelyconnectable, indefinitely repurposable, continuously updatable, and easily removed from thecontextofcollection”(MetcalfandCrawford,2016).

WhatisCrawfordandSchultz’S(2014)ProceduralDueProcessSolution?

In view of the limited protection offered by notice and consent rules during data collection,CrawfordandSchultz(2014)suggestedapplyingproceduraldueprocesstotheuseofdataandmetadataonindividuals.Thiswouldavoidexanteregulation(asoccurswhennoticeisgivenandconsentrequestedpriortocollection),insteadcontrollinganalyticalfairnessandequitythrougharbitration.63Suchaprocessrequirestherespondenttogivenoticeonhowdatawasprocessedandenablepartiestofilecomplaints.ThisapproachisbasedonDanielleCitron’sproposals(2010),inherarticle“TechnologicalDueProcess,”onautomatedgovernmentsystemsandtherisktheyposetofreedomandproperty.CrawfordandSchultzbuildonCitron’sideastoincludepredictivedataanalysis.

Thissystemwouldrequireentitiesengagedindataanalysistostatehowtheanalysesaretobeused, as well as to produce audit trails. Those involved in predictive data analysis would beresponsiblemustresponsiblefor“disclosingnotonlythetypeofpredictionstheyattempt,butalsothegeneralsourcesofdatathattheydrawuponasinputs,includingameanswherebythosewhosepersonaldataisincludedcanlearnofthatfact”(CrawfordandSchultz,2014,p.33).

63 There are seven enduring sets of values that due process should preserve: accuracy, appearance offairness, equality of inputs into the process; predictability, transparency and rationality; participation;revelationprivacy-dignity–eachofthesevaluesmapswelltoourconcernsaboutbigdata(CrawfordandSchultz,2014,p.19).


February2018


WhatOtherPossibleSolutionsHaveBeenIdentifiedbutnotExploredinDetail?

OtherIdentifiedPotentialSolutionsNotExploredinDepth

Thefollowingtablelistsotherpotentialsolutionsfoundintheliteraturereviewbutnotexaminedindepth(alignreferences).

Table2:PotentialSolutionsIdentifiedbutNotExplored


February2018


AppendixBPrivacyPrinciples

ExistingNorthAmericanprivacy laws are largelybasedon Fair InformationPracticePrinciples(FIPPs),governingthecollection,useanddisseminationofpersonaldata(RichardsandKing,2014;Schwartz,1999).Thesefiveprinciples,initiallyknownastheCodeofFairInformationPracticesCode,was established in 1973.64 They are often summarized by such terms as openness, uselimitation, individual participation (right to obtain/correct data), data quality and securitysafeguardsandconstitutethefoundationofUSprivacylegislation.65

Table3:BasicFIPPs

OECD Guidelines on the Protection of Privacy and Transborder Flows of Data are a secondcornerstoneofprivacyprotectionprinciplesandthebasisofmostWesternnationalandregionaldataprivacyregulations,especiallyinCanada(Cate,2006).Theseguidelinestakeanewapproachbyconsideringhowdata isused,andnotmerelyhowit iscollected.AsTable4 illustrates,thereasonsforusingdataandlimitationsonsuchuse,appearnexttotheFIPPs(transparency;limitsondatacollectionandaccess;monitoringthedata;qualityandsecurity).

64Aspartofthe1973reportentitledRecords,Computers,andtheRightsofCitizens,bytheUSgovernment’sAdvisoryCommitteeonAutomatedPersonalDataSystems.65AnotherkeyprivacydocumentwasproducedbythePrivateProtectionStudyCommissionduringJimmyCarter’spresidencyin1977.Wedonotdiscussthisdocumentindetailhere,sinceitshighlightsandthoseof the FIPPs were subsequently incorporated in theOECD Guidelines on the Protection of Privacy andTransborderFlowsofData.


February2018


Table4:OECDGuidelines

In1990,theEuropeanCommissionpublishedtheCouncilDirectiveontheProtectionofIndividualswithRegardtotheProcessingofPersonalDataandontheFreeMovementofSuchData,providinga roadmap to the adoption by EU members of national laws on the topic. It is particularlyinteresting that this directive focuses not just on personal data, but subsequent transfers ofcollecteddata(asforfuturethird-partyuse)andautomateddecision-making,twoissuesnotthencoveredbytheprinciplesdiscussedabove.ThefollowingtablesummarizestheDirective’scoreprinciples and two important rules for use—independent monitoring (outside audit of datamanagementanduse,andpersonal recourse ifharmed).TheDirective isconsideredthemostambitiousofallframeworksofthistype,todate(Cate,2006).


February2018


Table5:Principlesofthe1990EuropeanDirective

There isa final setofhighly influentialprinciples,whichpropose improvement in, rather thanreductionof, existing rules. These are theprivacyprinciplespublished in 1998by the FederalTradeCommission,anindependentUSagencyresponsibleforenforcingconsumerlegislationandmonitoring anticompetitive trading practices. These principles deal exclusively with onlineprivacy:


February2018


● Notice: Web sites must give consumers clear and prominent notice of their contentmanagementpractices.Thisisconsideredthemostbasicprinciple.

● Choiceandconsent:Websitesmust letuserschoosehowtheirpersonally identifiabledataisusedoutsideofthepurposeforwhichtheinformationwasoriginallyprovided.66

● Accessandparticipation:Websitesmustgiveusersreasonableaccesstoinformationthesitecollectsonthem,includingareasonableopportunitytoreview,correctanddeleteit.

● Integrityandsecurity:Websitesmusttakereasonablemeasurestoprotectthesecurityofinformationcollectedfromusers.

Manyobservershavementionedthatthelegislativeframeworkandcorporatepracticesoverthepastfewdecadeshavestronglyemphasizedtheconceptsofnoticeandconsent,ortheideathatusers/consumersshouldbecomeacquaintedwithandcheckthecontent-managementpracticesofcompanieswithwhichtheydealandallowtotransfertheirdatainexchangeforagivenservice.Thisprincipleisnowoftenaccompaniedbytheuser’sabilitytoconfiguretheirprivacyrules.

AnewGeneralDataProtectionRegulationwasadoptedinApril2016andhasbeeninforcesinceMay2018.Itisaimedatrestoringpeople’scontrolovertheirpersonaldata,whilesimplifyingthecorporateregulatoryenvironment.WhiletheRegulationreinforcescertainprinciples,itcontinuestorelyontheconsentprincipleandincorporatesthefollowingconcepts:

● Explicit,positiveconsent.● Righttoerasedata(ifpossible).● Righttopersonaldataportability.● Settinglimitstoautomatedforensicprofiling.● DefaultPrivacybyDesign.● Notificationofdatabreaches.● Appointmentbypublicandprivateorganizationsofadataprotectionofficer.● Mandatoryimpactassessmentsofanyactivitieswithpossibleprivacyimplications.● Encouragementindevelopingcodesofconduct(EuropeanParliament,2016;Wikipedia,

2017)

66However,as stated in its2012privacy report: companies shouldnotbecompelled toprovidechoicebeforecollectingandusingconsumerdataforpracticesthatareconsistentwiththecontextofatransactionorthecompany’srelationshipwiththeconsumer.ThisprincipleappliesequallytotheInternetofThings(FTC,2015,p.55)


February2018


AppendixCSolove’s(2007)andZiegeldorf’s(2014)PrivacyPrinciples

Manyof the activities involved in running anurban IoT can infringeonprivacy. Theworks ofZiegeldorf,etal.(2014)andofDanielSolove(2007)areparticularlyusefulinidentifyingpossibletension points—the first work concerns the smart city and the second, the general issue ofprivacy.

DanielSolove’sPrivacyClassifications

DanielSolove(2007)proposedprivacyclassificationsforidentifyingactivitiesthatarenowsociallyrecognizedasviolationsofprivacy,becausetheycanharmindividualsortheyarelikelytodosointhenearterm.67Aspresentedbelow,theseclassificationsareorganizedaroundfourstagesindatatransit:1) informationcollection,2) informationprocessing,3) informationdisseminationand4)invasion.68

67Solove’sclassificationschemehasbeengreatlyinfluencedbycurrentUScaselaw,butisalsoconcernedwithpossiblefutureinfringements.68Theauthorsbelievethatthenamesofsomeoftheseactivities,suchasinterrogationandexposurearenotalwaysintuitivelyclear,andinseveralinstancesquitedifferentfromexpressionscommonlyusedintheprivacyliterature.Theirpositionswithinthemodelarealsosometimessurprising.Theschemeremainsuseful,though,inacquiringabetterunderstandingofprivacyandissuesthatcouldinfringeonit.


February2018


It should be noted that some of these activities, such as interrogation and exposure (seedefinitions of those terms beneath Figure 7), have little or nothing to do with urban IoT.Furthermore,someactivitieswillnotbeadministeredbymunicipalagenciesinthecaseofurbanIoT,butbyexternalpartiesinteractingwiththeconnectedcity.

Figure7:DanielSolove’sClassificationScheme(2007)


February2018


Ziegeldorf’sDiagramofThreatstoPrivacy

AsappearsinFigure8,Ziegeldorf,etal.(2014)usedamodel69similartotheoneofSolove(2007),with two new features: 1) Interaction between sensors and people during initial datageneration.2)Disseminationofinformation,afterprocessingandanalysis,toindividuals.

Theirmodelidentifies:

● 4entities:sensors,people,datacollectioninfrastructureanddata-processingservices.● 5differentkindsofinformationstreams:initialinteractionbetweenpeopleandsensors;

information collection; data processing (analysis), generation of new information andservices;disseminationofinformationtothirdparties(includingpeople);presentationofservicescreatedtorecipients.70

● 7possiblethreatstoprivacy.

69ReferenceinspiredbytheTelecommunicationsWorkersUnion(TWU),aswellasthemodeldevelopedbytheEuropeanResearchClusterontheInternetofThings(IERC)70Ziegeldorf,etal.(2014)haveidentifiedothermodels,suchasiOT-iconsortium[18],Atzori,etal.[5],EUFP7projectsIOT-A[19]andCASGRAS[20].


February2018


Figure8:ModelofZiegeldorf,etal.(2014)71

71Greatereffortshouldbeappliedtounderstandingthetransitionlifecycleconceptfully.


February2018


AppendixD

SupplementtoSectiononSocialInclusionQuébec’s Statistics for Literacy, Numeracy and Problem Solving in Technology RichEnvironments(PSTREs)

There is a big digital divide inMontréal and throughoutQuébec. According to data from theInstitutdeStatistiquedeQuébec(2015),someoneinfiveQuebeckershavelowlevelsofliteracyandnumeracy.72Fifty-onepercentofthepopulationhasaverylimitedabilitytointeractwiththedigital environment.73While respondents of limited and very limited abilitywere still able tonavigatetheWebandperformsimpletasksonanelectronicdevice,theywerequicklystumpedbymorecomplexoperations,likebackingupdataandcompletingonlineforms.TheInstitutalsonotedthat17%ofsurveyrespondentsdidnothavetheirPSTREabilitiesassessedandprofilesofthat subgroup’s members suggest the large majority have low PSTRE scores. The study alsoreportedthat83.6%ofMontréalhouseholdswereconnectedtotheInternetin2012.74

Reference: http://www.stat.gouv.qc.ca/statistiques/science-technologie-innovation/utilisation-internet/menages-individus/menage-internet-2012.pdf

DiscriminationanditsRelationshiptoAlgorithmAnalysis

Weshoulddiscusstheconceptofdiscrimination, inadditiontothatof inclusion.Asweknow,stereotypes andprejudices are two characteristics of discrimination,which are both formsofcollectiveprofilingwhichoutweighreasoninginestablishingcharacterizationsofothers(Amossy,1989).Analgorithmis“discriminatory”ifitsintrinsiclogicincludesstereotypesandprejudices.Inthe social sciences, algorithms are “artifacts” associatedwith social practices. Algorithms areneverconsideredinandofthemselves(Hegel),butperceivedwithinacontextofthealgorithmandsocialpractices.

Stereotypesarecrude,oversimplifiedandrigidcharacterizationsofanobjectorgroup.Theyarecollective, pre-established, socially oriented and employed almost instinctively and routinely(Moscovici,2014).Theyaredecision-makinghabitsnotrootedinevidence,whicheverysocietyconveystoitsmembersthroughthefamily,socialcircles,schoolandmedia.Essentialthinking—orexplainingotherpeople’sconductandbehaviourintermsoftheir“essence”or“nature”—givesrisetostereotypes.

72“Low”means“inferior”andLevel1oftheISQevaluationscale.73 “Very low”means“inferior”and“low”meansLevel1of the ISQevaluation.PleaseseeAppendix fordetails.74http://www.stat.gouv.qc.ca/statistiques/science-technologie-innovation/utilisation-internet/menages-individus/menage-internet-2012.pdf


February2018


Ambrose Bierce75 defined prejudice, on the other hand, as “a vagrant opinionwithout visiblemeansofsupport,”or,moresimply,anunsupportedclaim.TheLaroussedefinesprejudiceas“apre-establishedjudgementaboutsomeoneorsomethingbasedonvariouspersonalcriteriaandthatpositivelyornegativelyinfluencefeelingstowardthispersonorthing,asinbeingprejudicedagainstsomeoneorhavinganunfoundedopinion,oftenimposedbysocietyorschool:Havingtheprejudicesofhisclass.”

Stereotypesandprejudicesareusedtocategorizepeople.Categorizationisamentalprocessoforganizingandstoringinformationonthelivingenvironment.Creatingcategoriesstructurestheworldandgives itmeaning. Theprocessofcategorization reliesona simplificationof reality,emphasizing the similarities between elements of one category and the differences betweencategories.

Categorization in social psychology is used in studying social relations. In algorithm analysis,categorizationisusedtoprofileapersonorasegmentofthepopulation.Thismeansstudyinghowencapsulationcreatesperceivedcategories,orputanotherway,thepartialhierarchiesthatalgorithmsestablishamongpeople,groupsandorganizations(BreyandSoraker,2009;Wiener1988).76Thisprocessmakesiteasiertounderstandthekindofnegativeorpositivebiasalgorithmstendtoproduce.

AntoinetteRouvroy77wrotethat,inthepast,statisticsweregeneratedthroughthecreationofcategories—categoriesarisingoutofacademic,technicalandpoliticaldiscourseandthetestingofideas,followingbydatacollectionsurveys.Now,however,shesaidthatcategoriescomeintobeingsimplybecausedataisavailable.Shebelievesthiswillnecessarilyleadtoschematizationoftheworld,withtheimminentemergenceofdigitalrealitygovernance.Inotherwords,wearechangingoursystemofgovernancebymakingwhatiscurrentlyvisibleobscure.

StrategiesintheLiteratureforPromotingPublicITEducationandAccess

Multipleobservershave recommendedpromotionofdigital education (Peres, 2015). France’sConseil économique, social et environmental (CESE) has proposed that its e-inclusion policyshouldbedevelopedaspartofabroad-based,ongoingpublicinitiative,andirrevocablylinkedtosocialinclusion.Theseproposalsinclude:

75TheDevil'sDictionaryisasatiricallexiconwithninety-ninedefinitionswrittenbyAmbroseBiercefrom1881to1906.76Brey,PandSoraker,JH,PhilosophyofComputingandInformationTechnology,Elsevier,2009;Wiener,N,Thehumanuseofhumanbeings:cyberneticsandsociety,DaCapoPress,1988.77AntoinetteRouvroy,HumanGenesandNeoliberalGovernance,Routledge-Cavendish,2007.


February2018


● PromotingaccesstobroadbandtechnologiesandInternet.● Supportingthefamily’sroleinteachingchildrenandyoungpeople.● Promotingcontinuouseducationinschool,fromkindergartenthroughgraduatestudies.● Encouraging the deployment/reinforcement of Wi-Fi access points and education for

adultsnolongerinschool.

Alladvancesindigitaleducationshouldnaturallygohand-in-handwithprogressinbasicliteracyand numeracy abilities. Peres (2015) suggested providing digital education on a variety oftechnologies, not just to acquire technical proficiency inusing such resources, but todeveloptrainingthatwillpromotethecriticaluseofsuchtechnologiesandeducationonpersonaldataprotection(Peres,2015).Inthesamevein,othershaverecommendednotfocusingontheuseofresources,butonenrichinginstrumentalabilities(handlinghardwareandinterfaces),aswellascreativeandproductiveskills(Conseilnationaldunumérique,2013).CESErecommendsbasicITeducation focusing on a variety of technologies that familiarize students with three basiccomputerconcepts:code,informationandalgorithm.

AlloftheseapproachesreflecttheConseilnationaldunumérique’srecommendation(2013)that“individualsshouldbecomeknowledgeableandresponsibleusersofdigitaldataandnotmereconsumersandthatuseofsuchdatanotbethesoleprerogativeofbusinessandgovernment”(ConseilNationalduNumérique,2013,p.5).

Support networks have been suggested for deploying/reinforcing Wi-Fi access points andeducation,sinceissuesofdigitalinclusionwillnowaffecttheentirepopulationandweareaimingatamovingtarget—thosecomfortablewithtoday’sdigitalsystemsmaybeoutoftheirdepth,tomorrow,asthesesystemsevolveandduetochangesinuseandgoalsofsuchuse.Furthermore,while digital functionality is generally growing, the digital realm is constantly expanding andincreasingincomplexity.However,thepeoplewhohavethegreatestneedsforsuchnetworksare those least likely to find a job, in a vulnerable situation or most disadvantaged (ConseilNationalduNumérique,2013).

Digitaleducationwillalsobenefitifitisnottreatedasa“catch-upact,”butaformofeducationcontributing to personal/social development and creativity. The digitalworld can help peoplerestoretheirself-esteem,eliminatetheirsocialexclusion,forgenewsocialnetworks,stimulatecreative behaviour, develop coordinated initiatives and facilitate democratic change (Conseilnationaldunumérique,2013).


February2018


AppendixESupplementtotheSocialInclusionSection

EthicalIssuesPosedbyAlgorithms,AccordingtoMittelstadt,etal.(2016)

Table6:ExtractofanArticlebyBrentDanielMittelstadt,etal.(2016)

Thesixethicalchallengesthatalgorithmsposetodecision-making

Lack of evidence orinconclusivereasoning

Thetermemployedinthedigitalsphereis“actionableinsight,”whichreferstoanintuition,mythorbeliefthatissufficientlyconvincingforthedecision-makertotakeactioneventhoughthecause-and-effect linkremainstobedemonstrated.

Incrementalevidence If data has been used (or produced) as evidence inmaking a statement.Normally, if the relationship between the data and conclusion is unclear,additional arguments should be developed. However, algorithms are notwell suited to the task. It ultimatelybecomes verydifficult to explain theoriginofdataanditsuseinproducingastatement.

Erroneousevidence Reliabilityisentirelydependentondataquality.Ifthedataiserroneous,theconclusionswillbe,too.

Unfairresults Actionstakenbasedonalgorithmicrecommendationsalsopossessanethicalcomponent.PleaseseeCOMPASsoftware.

Transformativeeffect Thealgorithmaltersmentalandsocialclassifications,creatingnewmeaning.This iswhywe refer to “algorithmic governance,” since itmodifies socialstructure.

Responsibility If a technology fails, penalties should apply proportional to the damagecaused.However, it isdifficult to identify thoseresponsible in thecaseofmachine-learningalgorithmsbecause“nobodyhasenoughcontroloverthemachine’s actions to be able to assume the responsibility for them”(Matthias,2004:177).78

78MatthiasA,“Theresponsibilitygap:Ascribingresponsibilityfortheactionsoflearningautomata,”EthicsandInformationTechnology6(3):pp.175-183,2004.


February2018


QuestionstheUSEPARecommendsAskingBeforePublishingEnvironmentalData

The US Environmental Protection Agency established a list of four questions to ask beforepublishingenvironmentaldata, topermitdatausers toevaluate itsqualityanddetermine if itcorrespondswiththepurposesforwhichitisbeingused(USEPA2006).Theyare:

● Canthedecisionorestimatebeproducedwiththedesiredlevelofcertaintyinviewofdataquality?

● Whatisthesurveyplan’sperformance?● Ifthesamesurveydesignstrategyisusedagainforasimilarstudy,shouldweexpectthe

datatosupportthesameuseswiththedesiredlevelofcertainty?● Haveenoughsamplesbeentakentopermitareviewertoseeaneffect if it isactually

present?(USEPA,2006)


February2018


AppendixFSupplementtotheSectiononFreedom

Lifelogging

“Lifelogging” is the accumulation of quantitative personal data pertaining tomultiple aspects(health,relationships,sportsperformance,etc.)ofaperson’slife.Varioussensorsandapps(suchasconnectedwatches),enablealifebloggertoacquire“self-knowledgethroughnumbers,”orinotherwords,personalstatisticss/hecanthenanalyzeandshare(Lupton,2016).Lifeloggingisadigitalactivityinvolvingautomated,continuousandcumulative“archiving”ofroutineactivitiesinthe form of digital data (images, graphics, geographic maps). This personal archive containsmultimodaldataobtainedthroughgenerallyubiquitousdigitaltools(sensorsplusapps).Thesesystems record all events, conversations, texts, audiovisual data, and traces generated bysociodigitalmedia,aswellasbiologicaldatageneratedbysensorswornonthebody,primarilytoprovideaccesstodataandthencross-matchitatafuturetime(Kelly,2007;DodgeandKitchin,2007).Bigdataproducedthroughlifeblogging,oftenwith“freeware,”canusuallybesold.

Subveillance

BaumanandLyonreferto“mini-panopticons”thatexistonapersonallevel.Inadditiontobeingmonitoredbythirdparties,everyonemonitorsthemselvesandothers.Ininteractingwithothers,each person plays a kind of supervisory role (by taking pictures, recording conversations andcapturing different data). “Subveillance” is the name occasionally used for such surveillanceconducted–sometimesunintentionally–bythepublicitself,usingpersonaldigitaldevices(Mann,2002)andcomplementsinstitutionalsurveillance(DodgeandKitchin,2007).Suchdataveillanceis all themorepervasivebecause it takes forms thathadneverpreviouslybeen suspectedofserving such purposes. Such examples as gamification (increasing a technology’s socialacceptability by applying typical elements of gameplaying to it), social interactions on digitalsocialnetworksanddigitalreputationmanagementcanentaildifferenttypesofautosurveillancearisingoutofcompliancewithvariousstandards.

PredictiveConsequenceAnalysis

Suchforecastsaredesignedtominimizeriskbypredictingtheconsequencesofaperson’sactions.Thiscouldinvolveencouragingindividualstobehaveinwaysbestcorrespondingtotheirinterests(often financial). For example, digital apps can be used to encourage a “poor” person not tobecomeanypoorerbyreducingpoordecision-making(Morozov,2015).


February2018


This first typeof forecasting isbasedonbehaviouraleconomicsandnudgingprinciplesand iscorrectlyfoundbysomeobserverstobeadverselypaternalistic(Morozov,2015).Aslaudableasthesepurposesmayseeminitially,consequencepredictioninfringesonpersonalautonomy,self-determinationandprivacy,andmorebroadlyresultsinthegovernment’ssheddingresponsibilityforpolicy,aswellas forsocialandcommunitydiscourse. Inotherwords, rather thanthinkingcollectivelyaboutproblemsandtheircausesintermsofcontextual,historical,socialandpoliticalissues,thisprocessisusuallyhandedovertobusinessesandminedforinformationwithoutregardtounderlyingcauses.


February2018


AppendixGSupplementtotheSectiononSocialAcceptability

“SocialAcceptability”DefinedintheLiterature

ThedefinitionofferedbyCaron-MalenfantandConraud(2009),authorsofthe“Guidepratiquede l’acceptabilité sociale : pistes de réflexion et d’action,” is frequently cited in Québec anddescribes social acceptability as “the result of a process inwhich the parties concernedworktogethertocreatetheminimumconditionsneededtoharmoniouslyintegrateaprogramorpolicyatagivenmoment,withinitsnaturalandhumanenvironment.”Thisdefinitionunderscoressocialacceptability’scoreprincipleanddemonstratesthatsocialacceptabilityisaco-constructionthatconsidersawiderangeofopinions.Itcanbeassumedthatthesolutionultimatelyimplementedistheresultofsomecompromise.

Beck(2001)proposedanotherwidelyuseddefinitionofsocialacceptabilityas“theanticipatedacceptanceofashort-andlong-termrisksrelatingtoaprojectorsituation.”Theconceptofriskassociatedwiththisdefinitionsuggeststhatthedegreeofsocialacceptabilitydependsdirectlyonwhatacommunitybelievesareacceptableprojectrisksintermsoftheirlikelinesstomaterialize(CPEQ,2015).

AccordingtoGendron(2014),in“Penserl’acceptabilitésociale:au-delàdel’intérêt,lesvaleurs,”social acceptability is the “public’s consent to a project or decision based on the collectivejudgmentthattheprojectordecisionissuperiortoknownalternatives,includingthestatusquo.”Thisdefinition,adaptedfromBrunsson(1996),focusesmoreonsocialacceptabilityastheresultof a collective judgment involving informed choices, including an excellent understanding ofpotentialopportunities,benefitsandrisks.

These considerations are also apparent in definitions by organizations working with majorprojects. BAPE, which is responsible for conducting public consultations, has defined socialacceptabilityas“asacollective,evolvingprocessthatbringsalargenumberoflocalandregionalstakeholders into play. It does not take the form of general consent, but of a consensus ofstakeholdersarisingoutofconsultationanddiscussion”(2014inBattelier,p.51),andassumesthatclientsanddecision-makerscanexplaintheirprojectsandhavethemapprovedbyormakethemacceptabletothepartiesconcerned.ThePMI’sWebsitesays,“Socialacceptabilityisavitalconcept,”because“someprojectselicitoppositionthattheprojectmanagementteammusttakeintoaccounttoensurethedesiredresultsareachieved.” Inthiscontext,socialacceptability isseenasafactorinthesuccessofmajorprojects.

final eport #1 for batch 5 of the iot standards

Documents