Collective Intelligence: An Overview

2

WhatisCollectiveIntelligence?TheInternethasgivenrisetosomeremarkabletechnologiesthatenablepeopletocollaborateenmasse. In barely a decade, the biggest encyclopaedia in human history has been written bymillionsofauthorswithnocentralisedcontrol. Inunderayear,ordinarypeopleclassifiedmorethan 50 million photos of distant worlds to help astronomers understand how galaxies areformed. All over the world, organisations are starting to open up social networks andcollaborative environments from which emanate an endless stream of data and insight. Thepotentialthatlieswithininterconnectednetworksofcomputersandhumansisonlyjuststartingtoberealised.

Weallpossessintelligence.However, intelligencecanbethoughtofnotonlyassomethingthatariseswithin human brains – it also arises in groups of people. This isCollective Intelligence –individualsactingtogethertocombinetheirknowledgeand insight.Collective Intelligence isanemergentproperty.Itemergesfromthegroup,itdoesnotresidewithinanyindividualmembers.

TheWisdom of Crowds is simply another term for Collective Intelligence, stemming from theworkofBritish scientist Sir FrancisGalton.Ata livestockexhibition in1906,Galtonobservedacompetitionwherepeoplehadtoguess theweightofanox.At theendof thecontest,Galtongatheredalltheguessesandcalculatedtheaverage.Thecrowd’sestimateturnedouttobenearperfect. Galton referred to this as the collectivewisdom of the crowd: the fact that groups ofpeople can be more intelligent than an intelligent individual and that groups do not alwaysrequireintelligentpeopletoreachasmartdecisionoroutcome.

Thetermcrowdsourcing issynonymouswithCollective Intelligenceandthetwoareoftenusedinterchangeably. Crowdsourcing can be defined in different ways, although it is frequentlydefinedas takinga task that is typicallyconductedbyonepersonandoutsourcing it toa largegroupofpeople. Crowdsourcing isusuallymoreofaprocess, thegoalofwhich is todistributeworkloadfromonetomany.CollectiveIntelligence,ontheotherhand,canbebetterthoughtofasagoal.Thefocusisontheoutput.

Within organisations, openness and transparency are becoming vital business characteristics.Similarly, considering society at large, the subjects of Digital Social Innovation and OpenDemocracyaregrowing inprominenceastheyaimtocapturehowcontemporarytechno-socialtrendscanbeharnessedtowardssolvingthemostimportantchallengesfacingmodernsocieties.

ThesechangesreflecttheriseofaneweraforCollectiveIntelligence,onethatisabletofullytakeadvantageofboththeincreasinginterconnectivitybetweenpeopleandtheincreasingcomputingand storage capabilities of advanced information and communication technologies. This is thefrontier of Social Collective Intelligence: networks of people and computers acting together inintelligentways.ThetermsocialmachinehasalsobeenusedtodescribeCollective Intelligencesystems and tools. It is helpful to think about Collective Intelligence emanating from socialmachinesbecause itemphasisesthejoint involvementofbothcomputersandthepeopleusingthemaspartofthemachinery.

3

UsingCollectiveIntelligenceTherearemanyareasofworkandtypesoftasktowhichCollectiveIntelligencemethodscanbeappliedandcategorised,withagreatdealofoverlapbetweenuses.

IdeationandInnovation

Collective Intelligence methods have been most commonly used for generating ideas andsolutions. Organisations apply Collective Intelligence tools with networks of employees,customers and other external parties. Many Collective Intelligence tools consist of virtualenvironmentswhereparticipantscaninteracttodiscussideasandopinionsorprovidefeedbackand evaluation on particular topics. Community ratings and commentary on participants’suggestions can be analysed, identifying themes and comments that resonate most with thecommunity. Shared interests can also bring people together in a community to enhanceknowledge,explorebestpracticeandproposenewideas.

KnowledgeSharingandWorkloadDistribution

In addition toallowingorganisations to gather ideasand feedback fromagreaternumberanddiversity of people, Collective Intelligence approaches also help organisations to act on thoseinsights by distributing work to people who are best placed to do it. This not only has thepotential to enhancequality, but increasing efficiency as tasks canbe conducted inparallel bymanypeopleatthesametime.

PredictingtheFuture

Aggregating diverse perspectives, knowledge and experience of employees and customers canimprove the accuracy of predictions, allowing organisations tomakemore informed, evidencebased decisions that can complement traditional forecasting approaches. Analysis can thendetermineforecastprobabilitiesforthevariousoutcomes.

Coordination&Collaboration

Coordination is when a community is createdwith the aim spreading information quickly andwidely.Increasingconnectivitymeansthatgroupsareabletocooperateandachievewhatoncewas the exclusive domain of large centralised organisations. Social media has the power torapidly organise and mobilise employees – even when they are geographically or temporallydispersed.

4

ChallengesforCollectiveIntelligence Although Collective Intelligence has been emergent throughout human evolution, the rise ofsocial technologiesover the last decadehas caused the rapiddevelopmentof awide rangeoftools, systems and platforms. Real-world implementations over this time have demonstratedsome of the obstacles that need to be addressed in the development and use of CollectiveIntelligencetools.

DesignChallenges

Researchers who are involved with the study of Human-Computer Interaction (HCI) andCollective Intelligence focus on how people and computers can be connected to enhanceusabilityandoutput.Thedesignchoicesmaderegardingtheuserinterfaceoftoolsareextremelyimportant as potential participants will only contribute if the user interface guides them in astraightforward and meaningful way. Similarly, those interested in leveraging CollectiveIntelligence require a means by which to administrate, analyse and extract insights from thesystem.Ultimately,CollectiveIntelligencerequiresthedesignofbothtechnicalinfrastructureandhuman-humaninteraction:asocio-technicalsystem.

List-baseddiscussionsandtheirevaluationSocialCollectiveIntelligencetoolsinvolvesomemeansofinteractionbetweenparticipants.Thistypicallytakestheformofanonlinediscussionorcommentssectionwhereparticipantsprovidetheir own input as well as explore, evaluate, or reply to, other's contributions. In onlinediscussions, lists of comments tend to hide the true diversity of opinion that exists, and canbecome impossible to navigate through increasing participation, quickly growing tooverwhelming proportions.Without ameans of effective navigation,many comments or ideaswill never be properly evaluated – participants do not have an equal chance of being heard.Collective Intelligence tools therefore need to be able to cope with large groups of people.Algorithms that share out comments are becoming increasingly sophisticated and can ensureeach contributionhasbeengivenanequal amountof attentionby the community. Interactivevisualisations also provide a compelling and engaging experience for participants who canthemselvesgetinstantandauthenticfeedback.

TimewastingInonlinediscussionsinvolvinglargenumbersofparticipants,thesameideasareoftenrepeated,sometimes preventing the discussion fromprogressing. Similarly, it can be difficult to organiseandcategorisemanydifferent ideas.Toomuchtimecanalsobewasteddiscussingsuggestionsthat areunworkableor impractical,meaning that good ideas can sometimesbeoverlookedorunderdeveloped.AggregationItcanbearguedthatthelevelofCollectiveIntelligenceprovidedbytraditionalsurveysislimited–thecruxofthematteristhemeansbywhichthedataisaggregated.Intelligencegainedfromtraditionalsurveyscanbeconsidered‘collective’inthesensethatthedataisbeingprovidedbymultiplepersonswithin thegroupand ‘emergent’,because the intelligenceproduceddoesnotresidewithinanymembersofthegroup.Advancementsinsocialtechnologieshaveallowednewmeans of aggregation to emerge. Groups can rate individual's written responses on specific

5

evaluationcriteria.Whentheseratingsarecombinedwithtextanalysis, it ispossibletoidentifywhichcommentsandthemesresonatemostwiththecommunity.Asaresult,outputisshaped,notbyanindividualanalystorteamofresearchers,butbymembersofthegroup.

AdministrativeChallenges

There are a number of administrative and operational hurdles to overcome in CollectiveIntelligence and there has been a great deal of research into understanding how activeparticipation in online communities can be enhanced and maintained. Overcoming thesechallengesiscrucialbecausehumaninputisfundamentaltoCollectiveIntelligence.

RecruitingparticipantsAt times, considerations about which demographics to recruit to participate in CollectiveIntelligenceactivitieswillbeneeded.Anycollective intelligence toolneeds tohaveaminimumnumberofactiveusers tohavemeaningful interactionsandmaintainengagement.Populationscanbeassmallas20andpotentiallyasbigaseveryonewhoisabletoread,writeandaccesstheInternet.Obtaininggoodparticipationratesisessential,ashavingarichdiversityofperspectivesisanimportantpre-requisiteforobtainingthebestCollectiveIntelligence.Itisimportanttoalsoreflect on whether participants require specific knowledge, understandings or experiences tocontributetoaparticularprojectorexercise.

MotivatingparticipantsandincentivisationParticipants need to be encouraged to share their input and interactwith others, all ofwhomhavevaryinglevelsofcommitment,expertiseandavailability.Participantsshouldbeclearwhatitistheyareapartof,whytheyshouldengageandwhatthevalueistoboththemselvesandtheorganisation. Intrinsic motivators such as social recognition, being heard (visibility ofcontributions),loyalty,havingacommonpurposeorbeinginterestedinthesubjectmatter,caninterplaywithmoreextrinsicmotivators,suchasvouchers,moneyorprizes.

LackofdetailinparticipantinputDepending on the specifics of the community and the topic being discussed, participants canoften provide input that is lacking in clarity or detail, with only a minority making carefullyconsidered responses.Of course, participants need to be able to answer the question or briefthey’regiven,butmoreimportantly,theyneedtofeelencouragedtocraftdeliberatedinputandtoreadandrespondtootherparticipants’input.

AnonymityAcommonproblemisthatmanyorganisationsdonotallowanonymisedcommentsintoolsandsystemsfromwhichCollectiveIntelligencecanbeelicited.Whendiscussingmanyorganisationalissues, what is required is honest feedback, which may be critical of the organisation or itsleadership.Many platforms in organisations do not allow for anonymised comments and, as aresult,meansthatemployeesareunlikelytobecandidintheirfeedback.

ConflictInanyonlinediscussion,therecanbesomeparticipantswhodonotsharethegeneralviewofthewidercommunityor,sadly,individualswhohavesetouttodeliberatelyabuseotherparticipantsordisruptthesystem.SomeCollectiveIntelligencesystemsaredesignedtoencouragepeopleto

6

read the comments of others who are different to them in some way, using visualisations toillustrate the diversity of opinion that is largely hidden in list-based discussions, or weightparticipants’ ratings of each other’s input, demonstrating various social, attitudinal ordemographic biases. A certain amount of conflict can be beneficial if it helps people tounderstandamorediverserangeofperspectives.

ManagerialandLeadershipChallenges

Collective Intelligence in organisations poses some specific challenges for management andleadership because tapping into Collective Intelligence means taking a more open andtransparentapproach–andthisrequiressignificantculturalchange.

LossofcontrolAsmoreworktakesonelementsofcrowdsourcing,movingtasksfromonetomany,thiscouldbeconstruedbysomemanagersasasignthattheyarenotdoingtheirjobsproperly.Seniorfiguresmay perceive the more open methods of taking perspectives and suggestions from acrossorganisationalboundariesandhierarchiesintoaccountasreducingtheirdecisionmakingpower.A common theme when talking about leadership in the digital era is that it necessitates adifferent way of doing things, but leaders often lack skills and awareness about CollectiveIntelligence tools, limiting their view of tangible organisational benefits and preventing themfrom driving change. Leadership must embrace change, be open to experimentation,demonstratetransparency,workcollaborativelyandpromotedialogue.

OvercomingresistanceOftenitcanbemiddlemanagerswhoaremostresistanttoamoreauthenticwayofoperatingorhesitantaboutembracinganopenapproach.Thisisperhapsbecausemanagersatthislevelaremost fearful about any critical feedback being attributed to them. In addition, it is frequentlymiddlemanagersthatowntheworkstreamstowhichCollectiveIntelligencecanbebestapplied.

IntellectualpropertyandsecurityissuesOrganisations can perceive the use of collective intelligence platforms as a risk to intellectualproperty, liabilityanddataprotectionissues,byallowingdiscussionsinanopenforum.Ifthis isthecase,agreementshouldbesoughtonpre-establishedtermsofparticipationandownershipofintellectual property. Another concernmight be open communication, inwhich large numbersparticipating in discussions on organisational issues could alert more people to potentialproblemsorsolutions.

LackofactioninresponseTheelementwhichorganisationsstruggletoaskforevaluativeresponsestoistheinstigationofchangeinresponsetoinsights,andprovidingfeedbacktoparticipantsonthefindingsandactionstaken.Inthisregard,CollectiveIntelligenceisnodifferent.Theinsightsmaybericher,morevalidandmoreactionable,andparticipantsmaybemoreengagedwiththeprocess,butunlessleadersand managers act on the insights, the Collective Intelligence obtained will remain redundant.Collecting participant feedback in an open, transparent and collaborative environment hasenormous potential for participants to actually learn and share information during datacollection.This can increase respondent’s readiness for change in that theyaremoreawareofthe issues at hand and are more likely to feel that their voices have been heard.

7

Conclusion

Therearemanyways inwhichorganisationsareusingCollective Intelligence. From ideation toaugmenting skills and predicting the future, an increasing number of tools are allowingorganisations to more easily address critical issues. Whatever the specifics of the approach,CollectiveIntelligenceisenablingorganisationstotapintothesharedknowledgeandexpertiseoflargeanddiversegroupstoaddresstheirmostimportantorganisationalchallenges.

Successful organisations have been encouraging collaboration and interaction since the firstmodernorganisationswereformalised.ItisSocialCollectiveIntelligence,resultingfromtherapiddevelopment of social and digital technologies, which is both new and exciting in this area.Organisationsareonly justbeginningtoappreciatethepotentialthatCollectiveIntelligencecanoffer.Therearepocketsofgoodpracticeappearingasorganisationsstartto implementvariousCollective Intelligenceactivities.However, theseearly examplesof Social Collective Intelligencearelikelytobejustthestartofthestory.

As theareaprogressesandevolves, thosewhoaredeveloping systemsand tools forCollectiveIntelligence will need to take into account the various challenges and pitfalls outlined in thisreport.Notonlyinthedesignofhuman-computerinteractions,butalsotakingintoaccountthepsychologyofsocialandcognitiveprocessesthatoccurwhengroupsofpeopleinteractonalargescale. Moreover, the way in which Collective Intelligence systems are implemented andembeddedwithinorganisationsneedstobeconsideredalongsidetheadministrative,managerialandleadershipchallengesthatsuchapproachespresent.

SimplesurveymethodologiesdoprovideCollectiveIntelligenceofasort,butwhattheylackisasocial element that enables participants to evaluate each other’s responses. If a group isinteracting in a system that has been carefully designed to optimise their input and extractinsightthenthelevelofCollectiveIntelligenceobtainedcanbefargreaterthananysurvey.

Theprogressbeingmadewithinthefieldsofartificial intelligenceandmachinelearningis likelyto be at the forefront of Collective Intelligence as it evolves in the coming decades. At themoment, it is typically the humans that do the ‘intelligent’ work and computers that do theadministration.ThinkingaboutthefutureofCollectiveIntelligence,itisclearthatdevelopmentsinhuman-computerinteractionarelikelytoplayabigroleenhancinghuman-humanandhuman-machine interaction. For many scientists and entrepreneurs, the ultimate aim is to develophybridsystemsthatemploybothhumanintelligenceandadvancedmachineintelligence.

8

©CopyrightCrowdoscopeLtd2015CrowdoscopeArch462,KingslandViaduct83RivingtonStreetLondonEC2A3AYwww.crowdoscope.com

Crowdoscope isa tool forobtaining thecollective intelligenceof smallor largegroupsofpeople.It isaself-organisingvisualenvironmentthatcansupportanunlimitedamountofparticipants.InCrowdoscope,groupscaninteractinasystemcarefullydesignedtooptimisetheir input and extract insight. In this way, Crowdoscope helps communities navigatediscussionsmoreeasily,facilitatescollaborationandenhancesparticipantinvolvement.