Learning Center
Plans & pricing Sign in
Sign Out



 U-II   98.18-23           APRIL

        Persuasive Computers:
                                      1998                                                                                                        PAPERS

                                 Perspectives                            and Research Directions
                                                                                  BJ Fogg

                    Stanford University                                                                   Sun Microsystems
            Box 8333, Stanford, CA 94309 USA                                                      901 San Antonio Road, MPK17-105
 ,                                                        Palo Alto, CA 94303 USA

   The study of computers as persuasive technologies (called
   “captology”) was introduced at CHI 97 as a new area of
   inquiry. This paper proposes definitions, perspectives, and
   researchdirections for further investigation of this field A
   persuasivecomputeris an interactive technology that attempts
   to changeattitudes or behaviors in some way. Perspective1
   describes how computers can inherit three types of
   intentionality: endogenous, exogenous, and autogenous.
   Perspective 2 presents the ‘Functional Triad,” which
   illustrates that computers can function as persuasive tools,                         Figure 1: Captology describes the shaded area   where   computing
   media, and social actors. Perspective3 presents a ‘levels of                         technology and persuasion overlap.
   analysis” approachfor captology, which include varying levels
   from individual to societal- Perspective4 suggests a simple
   method for exploring the design space for persuasive                                 This paperfirst presents five perspectiveson computers and
   computers. Perspective 5 highlights some ethical issues                              persuasion.They include the following:
   inherent in persuasive computing. The paper concludes by                                           1:
                                                                                          Perspective Definition of persuasive   computers
   proposingsevendirectionsfor further researchanddesign.                                            2:
                                                                                          Perspective A functionalview of persuasive  computers
                                                                                          Perspective Levels of analysisfor captology
                                                                                                     4:                   for
                                                                                          Perspective The designspace persuasive       technologies
   persuasion, captology, media, computers as social actors,
   ethics,designmethods,computersaspersuasivetechnologies                                            5:                     that
                                                                                          Perspective Ethicsof computers persuade
                                                                                        Each of these five perspectivesprovides a different way to
   INTRODUCTION                                                                         view persuasivecomputers,while also describingexamplesof
   At CHI 97, a specialinterest group meeting gathereda number                          relevant technologies.After setting forth the five perspectives
   of participantswho were interestedin exploring the domain of                         on persuasivetechnologies,this paperconcludesby outlining
   computersand persuasion [a- We agreedto call this area                               sevenprofitable directions for further researchin the area of
   “captology” (built from an acronym for Computers As                                  persuasivecomputers.
   Persuasive Technologies), with the graphic in Figure 1                               To be clear, this papermakes a contribution by articulating a
   serving as a referencepoint for this domain.                                                              to
                                                                                        range of approaches captology. It is the role of later work to
   The discussionon captology at the CHI 97 SIG proved fruitful                         expandandrevisethe ideasproposedin this paper.
   and enlightening, with participants concurring that captology
   was an intriguing area for further researchand design. The                           PERSPECTIVE   1:
   group also agreedthat this domain had not yet been adequately                        DEFINITION  OF PERSUASIVE               COMPUTERS
                        by              and
   definedor addressed researchers practitioners of human-
   computerinteraction. We found that our discussionsufferedat                          What   is a persuasive      computer?
   times because lacked key definitions and frameworksfor                               Simply put, a persuasive computer is an interactive
   understandingcaptology.The purpose of this paper, therefore,                         technology that changesa person’sattitudesor behaviors.This
   is to contribute to the CHI community’s understandingof                              definition works well in many cases,but a more thorough
   persuasivecomputing by proposing definitions, perspectives,                          definition gives a better understanding of persuasive
   andresearch   directionsfor the field of captology.                                  computing.
   l’cnnission to make digimlkurl copies ofzdl or pti of this msterial for              The psychology literature suggestsmany definitions for the
   pxsonal or clssroom use is gmnted witbout fee provided that the copies               word ‘persuasion” [e.g., 24, 361. After reviewing the work of
   are not made or diibuted for profit or commcrcird advantqe, the copy-                persuasionscholars,I’ve synthesizedthe various definitions to
   rightnotic+tbetitleoftbepubliwtionand          its&teappear,andnoticek
   given that copyright is by permission oftbeACh% Inc. To copy otbentise,
                                                                                        define “persuasion” as “an attempt to shape, reinforce, or
   lo rcpubliih, lo post,on servers or to redistribute to Iii requires -qecific         changebehaviors,feelings, or thoughts about an issue, object,
   p~mtission andlorfez
   CHI 9SLosAngeles USA
   copyright o-g9791-975-o19W 4.s.00
                     CA                                       This is trial version     or action.”

                          -   -                 .A.    -^   _   -                --.-   ,-          . . .   .ii   ;   ^r-.--   -...‘---x^-..      _-,.   ‘.&:..-,..   -,_    I_-,
   Persuasion and intentionality
                                                                                                                                     CHI       98 . 18-23

                                                                                             Furthermore, it is quite possible that a given interactive
                                                                                             technologymay fall into more than one category.Despite the
                                                                                                                                                                            APRIL   1998

   One key point in this definition is that true persuasionimplies
   an intent to changeattitudes or behaviors; in other words,                                potential ambiguities, I find thesethree categorieshelpful in
   persuasionrequiresintentionality. Therefore,not all behavior                              better understanding the range and roles of persuasive
   or attitude changeis the result of persuasion.For example,a                               computing technologies.
   rain stormmay causepeople to buy umbrellas, but the storm
   is not a persuasiveevent becauseit has no intentional@                                    PERSPECTIVE 2:
   associated  with it. (However, if an umbrella manufacturer                                A FUNCTIONAL VIEW OF PERSUASIVE
   could somehowcauserain, then the rain storm might qualify                                 COMPUTERS
   asa persuasive  tactic.)                                                                  While Perspective1 provides an “intentional” framework for
   Becausemachines do not have intentions [S], a computer                                    persuasivecomputers, Perspective2 presents what I call a
   qualifies as a persuasivetechnology only when those who                                   “functional” view. To explain Perspective2 clearly, I first
   create,distribute, or adoptthe technologydo so with an intent                             describethe basicsof the “Functional Triad.” I then show how
   to affect human attitudes or behaviors. To be clear, the                                  this framework can provide key insights into the study of
   persuasive  natureof a computerdoesnot residewith the object                              persuasive computers.
   itself; instead, a computer being classified as “persuasive”
   depends the context of creation, distribution, and adoption.
             on                                                                              The Functional Triad
   I proposethat if an intent to changeattitudesor behaviorsis a                             I proposethat today’s computersfunction in three basic ways:
   factor in the creation, distribution, or adoption of a                                    as tooIs, as media, and as social actors. Researchers     and
   technology,then that technology inherits a type of intent from                            designers  have often discussedvariantsof thesefunctions[e.g.,
   humanactors.                                                                              18,22,33], usually as metaphorsfor computer use. However,
                                                                                             I suggestthat these three categoriesare more than metaphors;
   Threetypes of intent- endogenous,exogenous,and                                            they are basicways that peopleview or respondto computing
   aufogenous                                                                                               I
                                                                                             technologies. refer to thesethreeareasas “functions.”
   For the purposes of captology, I propose three kinds of                                   As a tool, the computer (or the computer application or
   inherited persuasive intent: endogenous, exogenous, and                                   system)provides humanswith new ability or power, allowing
   autogenous. According to my definitions, a computer                                       peopleto do things they could not do before,or to do things
   technology inherits endogenous intent when a designeror                                   more easily [28,29].
   producer       a                                  usersin
            creates technology with intent to persuade
   some way. A computer technology inherits exogenous                                        Computersalso function as media [13, 30, 321, a role that
   intent when one person provides another person with a                                     has becomemore apparentand important in recentyears [30,
   computer technology in an attempt to changethat person’s                                  341. As a medium, a computer can convey either symbolic
   attitudes or behaviors. A computer technology inherits                                    content(e.g., text, data graphs,icons) or sensorycontent(e.g.,
   autogenous intent when a person choosesto use or adopt a                                  real-time video, simulations,virtual worlds).
   technology in order to changehis or her own attitudes or                                  Computerscan also function as social actors [12, 16, 25,
   behaviors.Table 1 makesthis idea clearer.                                                 321.Usersseemto respondto computersas social actors [25]
                                                                                             when computer technologies adopt animate characteristics
                                                                                             (physical features, emotions, voice communication), play
   I Type of intent      Where intent comes
                                                      Example                                animate roles (coach, pet, assistant, opponent), or follow
                                                                                             social rules or dynamics(greetings,apologies,turn taking).
                         Those who create or          Health-Hero video games are
                         produce the                  designed to persuade                   Mappingthe funcfions
                         interactive                  children to develop good
                         technology                   health habits [IT&
                                                                                             One way to view thesethree functions simultaneously is to
                                                                                             map the three categories into two dimensions. Of course, in
                         Those who give               A mother may give her son a            all but the most extremecases,thesefunctions mix and blur
                         access to or                 Palm Pilot PDA in hopes that           in any one given interactive technology. However,
                         distribute the               he will become more
     external factors’                                                                       consciouslymappinginteractive technologiesinto a triangular
                         interactive                  organized.
                         technology to others
                                                                                                                                        insight into the
                                                                                             spaceI call the ‘Functional Triad” generates
                                                                                             roles and relationships of different computer technologies.
     autogenous          The person adopting          A person may by and use a              Figure 2 represents the Functional Triad with some
                         or using the                 calodecounting computer                prototypical examples.
                         interactive                  device to help change his or
                         technology                   her own eating behavior.
                                                                                             Persuasion and the functional view of computers
   Table 1: Three types of intent wfth   examples                                            Although the functional triad is a useful framework for
                                                                                             understanding computertechnologiesin general,adoptingthis
   Although the abovecategoriesaim to distinguish amongtypes                                 functional view of computers yields specific insights for
   of persuasivetechnologies, recognizethat thesecategories
                            I                               are                              analyzingpersuasiveinteractive technologies.
   not always precise, and they are not always mutually
   exclusive.Indeed,making inferencesaboutintentions is tricky
   business-we may infer intent where thereis none, or we may
   fail to perceive intent when intent does indeed exist.
                                                            This is trial version
                                                                               Examples     in each functional     area
                                                                               As the Table 2 shows,computersfunctioning as tools, media,
                                                                               or social actors can change attitudes and behaviors through
                                                                               different means. Examples of persuasivecomputers in each

                                                                               Exampleof computeras persuasive tool
                                                                               Polar Heart Rate Monitor []: This exercise
                                                                               devicesoundsan alarm when a person’sheart rate falls outside
                                                          Social               a pre-setzone. The devicenot only can motivate a person to
                                                          Actor                changebehavior during exercise,but it may also increaseself-
                                                                               efficacy about one’s ability to exercise effectively (thus
                                                                               increasinglikelihood of adherence an exerciseprogram).

                                                                               Eirample of computer as persuasive medium
                                                                               HIV Roulette []: A computerized
                                                                               exhibit at the San Francisco Exploratorium allows visitors to
   Medium                                                                      make hypothetical choices about sexual behavior and then
                                                                               vicariously experiencehow those choices would affect their
 Figure 2: The Functiond   Triad with examples                                 chancesof contracting HIV. This exhibit attemptsto motivate
 By viewing a computertechnology as a tool, one can then ask                   peopleto avoid unsafesexual behaviors.
 how tools canbe persuasivedevices.In other words, ‘KOW do
 tools change attitudes or behaviors?” While this question                     Example of computeras persuasive social actor
 deserves more detailedexploration, one preliminary answeris                   “5-A-Day Adventures” []:          This CD-
 that tools can be persuasive by (1) reducing barriers and                     ROM, designedfor children, features a characternamed HB
 thereby increasede likelihood of a certain behavior [2, 111,                  who guides users through the application. HB encourages
 (2) increasingself-efficacyby making a certain behavior seem                  usersto eatmore fruits and vegetables,and he praisesthem for
 achievable [2, 171, (3) providing information that allows                     compliance.
 informed ,decisions[19], and (4) shaping a person’s mental
 model by channelingbehavior in a certainpattem-                               Functional    view   illuminates    affordances
 One could also pose similar questions about the other two                     The functional view of persuasive computers offers key
 functional areasfor computers: ‘What makes computers as                                                              of
                                                                               insights into the different affordances persuasivecomputers.
 media persuasive? and ‘What makes computers as social                         Not only does this framework help differentiate among
 actorspersuasive?”[4,5,20]. While this paper will not fully                   persuasivecomputers(as well as among different aspectsof a
 answerthese questions, Table 2 highlights some persuasive                     single persuasivetechnology), but this perspectivecan also
 affordances eachof the threefunctional areas:                                 provide researchers    with guidancefor primary and secondary
                                                                               research  efforts.
   Function          Essence            Persuasive affordances
                                                                               PERSPECTIVE 3:
   computer as       increases          l    reduces barriers (time, effort,   LEVELS OF ANALYSIS            FOR CAPTOLOGY
   Pool or           capabilities            cost)
   instrument                                                                                       I
                                                                               The third perspective proposefor understandingcomputersas
                                        l    increases self-efficacy           persuasivetechnologiesis one that directs specific attention to
                                        l    provides information for better   various levels of analysis for both computer use and
                                             decision making                   behavior/attitudechange.
                                        l    changes mental models
   computer as       provides           l    provides ffrst-hand learning,     Levels   of analysis:    Not just   individuals
   medium            experiences             insight, visualization, resolve   Most people think of computer use as it applies to an
                                        l    promotes understandiig of         individual-an individual interacts with or through a
                                             cause/effect relationships        computer.The sameholds true for notions of persuasion:An
                                                                               individual changeshis or her attitudes or behaviors. In both
                                        l    motivates through experience,
                                                                               cases,this level of analysis focuses on the individual. Even
                                                                               though both computeruseand persuasionare often thought to
   computer as        creates           l    establishes social norms          be most germane to individuals, other levels of analysis
   social actor       relationship      l    invokes social rules and          besidesthe individual level can generateimportant insights in
                                             dynamics                          researchingand designingcomputing technologies.
                                            . provides social support or       In recent years, HCI researchers have done admirablework in
                                              sanction                         conceptualizingcomputeruse on a level of analysis other than
 Table 2: Three computer functions and their persuasive affordances            individual [see reference 10 for resources; see also
                                                           This is trial version
                                                                     ].         Usually referred to as

    “computer-supported    cooperative wore’ this area examines
    computing asit pertainsto groups or organizations(group and
                                                                                                                  CHI    98   l   18-23   APRIL   1998

                                                                                    The table showsthat, at least in principle, certain technologies
                                                                                     are best suited for different levels of analysis-some are
    organizationallevels of analysis). It is important for the HCI                  individual-level technologies and some are community-level
    community to continue defining and exploring the various                        technologies. (Admittedly, today’s default view is biased
    levels of analysisrelatedto computeruse.                                        toward computers as individual-level devices; however,
    The fields of public information campaigns and health                                      in
                                                                                    advances online technology are making community and
    promotion interventions have also done notable work in                          societal interactive technologies more common. To be sure,
    conceptualizingdifferent levels of analysis [e.g., see 11. The                  the rise of the Internet has helped shift our focus to larger
    levels of analysisfrom thesefields include-                                     levels of analysis.) Furthermore, the table implies that
        * intraindividuallevel                                                      behavior change is not just an individual-level variable:
        l  interindividuallevel (dyads, couples, friends)                           families change, organizations change, and communities
            l    family level
            l    group level                                                        Although this section of the paperonly begins to explore the
                                                                                    benefits of conceptualizingpersuasivecomputersat different
            l    organizational level
                                                                                    levels of analysis, the potential benefit to HCI and captology
            l    community level                                                    is substantial. Public information campaign researchers    have
            l    societal level                                                     accomplished much by conceptualizing interventions at
    To be sure,previousHCI work hasexaminedcomputing at all                         different levels of analysis[ 1,261.The sameshouldbe true for
    of the abovelevels, but our field can benefit from more clearly                 those of us working in HCI-related fields, especially those
    differentiatingthesedifferent levels andfrom deliberatelyusing                                            or
                                                                                    interestedin researching designing interactive technologies
    theselevels more often as a driving force in researchingor                      that promotebeneficial social changes[27].
    designing computing technologies. HCI and captology have
    much to gain from viewing computing from these various                          PERSPECTIVE  4:
    levels of analysis.                                                             THE DESIGN SPACE          FOR       PERSUASIVE
    Different levels of analysis cause different variables to be
    salient, which generates   new insights in researchor design.                   The previousthreeperspectives  proposeframeworksfor better
    The different levels also draw on different theoretical                         understanding persuasivecomputing technologies. Although
    frameworks, methodologies, and measures. For example,                           the ideasfrom the previous sections serveas analytical tools
    persuasion on the individual level is conceptualizedand                         to study existing persuasivecomputing technologies, these
    measured  differently than persuasionon the community level.                    frameworkscan also serve a useful generative    purpose. In
    Although examining each level as it applies to persuasive                       otherwords,the previousframeworkscan help createideasfor
    computersis beyondthe scope of this paper, Table 3 gives                        new types of persuasivecomputing technologies.When used
    someexamplesto help illustrate this point.                                      for their generativepowers, the three previous perspectives
                                                                                    contribute to my fourth perspective: the design space for
                                                                                    persuasive technologies.
        Level of analysis    Computer artffacb             Behavior change of       The captologydesignspaceis large and relatively unexplored.
                             application, or system        interest
                                                                                    Because  captology presentsa new perspectiveon the role of
        individual           Baby Think It Over            To motivate an           computers, it is relatively easy to conceptualize new
                             (A computerized doll that     individual to be more                                                    to
                                                                                    technologies,or to add persuasiveenhancements existing
                             requires care. Designed to    sexually responsible     technologies.
                             teach teens about the         so he or she doesn’t
                                                           become a teenage         To approachthis design spacemethodologically, I propose a
                             diiculties of parenting                                two-step process:(1) identify domainsand issues of interest,
                             [                parent.
                                                                                    and (2) usethe aboveframeworksto gain new perspectives.
        family               Home video game system        To increase family
                                                           interactions (possibly
                             (A parent may want her                                 Identifying   domains   and   issues
                             family to interact more       measured by number
                                                           of conversations,        My searchesin the academicliterature indicate that most
                             wilh each other rather than
                                                           frequency of eye         existing computersystemswhich attempt to changeattitudes
                             passively watching TV.)
                                                           contact, etc.)           or behaviors fall into just a few categories: health, mental
        organixational       Remote work system            To reduce
                                                                                    health, and education [e.g., see 311. My efforts to identify
                                                           absenteeism.             persuasive                                              also
                                                                                                interactivetechnologiesin today’s marketplace
                             (A computer system that                                showedhealth as a central domain. Of course, health and
                             allows people to work         To increase              educationare still excellent areasfor researchand design of
                             effectively from home.)       employee retention.
                                                                                    persuasive technologies, but many other domains remain
        community            Ride-sharing system           To reduce the            relatively unexplored.
                                                           community% use of
                             (A community paging
                                                           private cars.
                                                                                    When my students and I set out to identify appropriate
                             nehvork that allows people                             domains for captology, we first made an extensive list of
                             to coordinate ride sharing    To reducetraffic         pervasivesocial or personalproblems. We then clusteredthe
                             with minimal prior            congestion and
                                                                                    problems and nested them under broaderheadings. In this
                             planning.)                    pollution.
                                                                                    exercisewe developed long (though not exhaustive) list of
   Table 3: Levels of anal@, tecbnologi~~,and behavioral outcc~mes                  domainsin which persuasivecomputertechnologiesmay serve
                                                       This is trial version
 a useful role. Table 4 contains a brief excerpt from our list       The computeracting as a medium could then renderan image
 (seewww.captology.orgfor more domains).                                                                  to
                                                                     of a tree growing as paperis added the recycling container.
                                                                     (The image might be on the recycling container itself.) Seeing
 I-                                                                  the tree grow in size might provide motivation to recycle
      Domain                 Issues
                                                                     because the image helps people visualize a cause/effect
      safety                 .   safe driving                        relationship in recycling.
                             l   wearing bike helmets                The computer as social actor could audibly thank people
                             l   substance abuse                     eachtime they addedpaper to the recycling container.A more
      endronment             l   recycling                           extreme idea is to have an anthropomorphictree that smiles or
                             .   conservation                        tells humorousstories in responseto recycling efforts.
                             l   bicycle commuting                   Intentionality      perspective: The recycling technology
      personal management    .   time management                     described  above would inherit both endogenous persuasive
                             .   study habits                        intent (from the designers of the device) and exogenous
                             .   personal finance                    persuasiveintent (from the management     teamthat providesthe
 Figure 4: A sample domains persuasive computers
                  of      for                                        technology for rest of the organization).
                                                                     To further explore the designspace,one might profitably ask,
 Using captology frameworks                                          “How might au autogenously              persuasive technology
 Once a person selects a domain or issue, the next step in           function in this organizational setting?” Becausethis design
 ,exploringthe design spacefor persuasivecomputersis to use          examplefocuseson the organizationallevel, autogenous     intent
 the captologyframeworksdescribedin this paper to generatea          would imply that the organization as a whole (not just the
 variety of perspectivesand ideas about the selected domain or       management)    would chooseto adopt a technology to change
 issue.For example,onemight ask the following questions:             their collective recycling behavior.
 Levels of analysis: What do different levels of analysis
                                                                     Design explorations      generate insight
 offer to this issue or problem? Which level of intervention is
 likely to be most productive?                                       Exploring the design spacefor persuasivecomputersis often
                                                                     an enlightening process.Especially if one has no intentions of
 Functional       Triad: What functional areas will the              turning ideas into marketable products, one can push the
 persuasivetechnology leverage?Should it focus on computer           envelope of the design space to illuminate new theoretical.
 as tool, medium, social actor, or combinations of the three?        areasof persuasivecomputing. As the above exampleshows,
 Inteutionality:      For this issue or domain, what are the                                                              and
                                                                     the two-stepmethod of (1) identifying domains/issues then
 possibilities for endogenously,exogenously, or autogenously         (2) applying the captology frameworks is a simple process
 persuasive  interactive technologies?                               that yields rich-and often surprising-results.
 An exampleof using thesefiameworks in design exploration
  follows.                                                           PERSPECTIVE  5:
                                                                     ETHICS OF COMPUTERS THAT PERSUADE
  Example: Recycling paper products                                  Ethics is yet another perspective from which to view
  For this example,I have chosen recycling as the issue to           computers as persuasive technologies. Adopting an ethical
  address using persuasive technology. Specifically, this            perspective on this domain is vital because the topic of
  examplefocuseson recycling paperproducts.                          computers and the topic of persuasionboth raise important
                                                                     issuesabout ethics and values [7, 15,361.Therefore,when the
  Levels of analysis perspective: One can profitably                 domains of computers and persuasionmerge (as they do in
  conceptualizerecycling paper at various levels: individual,        captology), the ethical issues are sure to play a crucial role.
  family, organization, or community. Of these levels, I have        This sectionpresentsexamplesof some ethically questionable
  chosen the organizational level for this example. The goal,        technologies,proposesethical responsibilities for designersof
  then, is to explore the possibilities for creating a persuasive                                                     and
                                                                     persuasivecomputersand captology researchers, discusses
  technology for increasing paper recycling behavior at the          the importanceof educatingaboutpersuasion.
  organizational level. (At this point one might then study
  various issuesfor more insight: organizationalculture, barriers    Persuasive    technologies     that   raise   ethical
  to recycling, etc.)                                                questions
  Functional Triad perspective: One would also explore               Most examples of persuasive computers in this paper ate
  how the three different functions of computers(tools, media,       technologies that promote widely held conceptions of the
  social actors) might profitably motivate paper recycling           “good” [23]: a computerizeddoll motivates responsiblesexual
  behavior. Even on the organizational level, many ideas are         behavior,a CD-ROM encourageseating fruits and vegetables,
  possible.For this example, one might proposethe following:         a ride-sharing technology cuts down on pollution in a city.
  A computeracting asa tool could weigh the recycled paper in        However, persuasive technologies can also serve ignoble
  the organization and calculatethe number of treessavedfrom         purposes.Table 5 contains examplesof persuasivecomputing
  recycling efforts. Other types of calculations are also possible   that may be ethically questionable, along with a brief
  (trees savedper week, oxygen prodncedby the saved trees,           gain/loss analysis for stakeholders (which is, admittedly,
  erc.). Knowing this information might motivate people in the       highly subjective).
  organizationto recycle their paperproductsmore.   This is trial version

    Summary of persuasive               Stakeholder analysis
                                                                                                               CHI   98    . 18-23

                                                                                the effects (the unintended side-effects) of persuasive
                                                                                                                                       APRIL   1998

    technology in question                                                      technologiesor strategies.
    A computertzed slot machine         gain: manufacturer and casinos          3. Disclose     findings.    Those who study persuasive
    uses animation and narration        loss: indiiduals (money, time)
                                                                                computersthen have an obligation to disclosetheir findings.
    to make the gambling
                                                                                4. If needed, take or advocate social action. Finally,
    experience more compelling.
                                                                                if a computing artifact is deemed  harmful or questionablein
    A computer system records           gain: company                           someregard,a researcher   should then either take social action
    employees’ activities on the        loss: employees (privacy, trust)                   that
                                                                                or advocate othersdo so.
    A computer system monitors          gain: restaurant patrons                Education    is the   key
    restaurant employees’ hand
    washing behavior after using
                                        loss: employees (privacy, trust)        I propose that the best approach regarding the ethics of
    the restroom.
                                                                                computersandpersuasion to educate      widely about this new
                                                                                areaof research design. Education about captology helps
     A software installation            gain: company                           people in two important ways: First, increasedknowledge
     seems to require registration      loss: indffidual users (personal        about persuasivecomputers allows people more opportunity
     with the company in order to       information)                            to adoptsuchtechnologiesto enhancetheir own lives, if they
     complete the installation.
     The software automatically
                                                                                choose.Second,knowledgeabout persuasivecomputershelps
     dia1.scompany to dovmload                                                  people recognize when technologies are using tactics to
     personal information.                                                      persuade them. In the end, the best path for captology is the
                                                                                sameas the path taken by various persuasionresearchers15,[
   Table 5: A simplestakeholder analysis of somepersuasive technologies
                                                                                24, 361: Educate widely about persuasion. This paper
                                                                                represents effort to bring such issuesto light.
   I believe that the simple gaitioss analysis in Table 5 helps
   show why the above technologies could be ethically                           RESEARCH       DIRECTIONS          FOR     CAPTOLOGY
   questionable. most casesabove, companiesstand to gain                        Because captology is a relatively new field of inquiry, many
   profit or information, while individuals standto lose money,                 questionsaboutcomputersand persuasionremain unanswered.
   privacy, or freedom.In contrast, a simple gain/loss analysis                 To help move the field forward, I proposesevendirectionsfor
   for the other persuasivecomputersin this paperwould likely                           and
                                                                                research designthat will yield the greatest  understandingof
   show gainsfor all stakeholders, including individuals.                       persuasivecomputersin the shortestamountof time.

   Ethics of distributing              or creating       computers         to   Seven   profitable    directions     for   captology
   change  attitudes     or           behaviors
   The ethical implications for those who design persuasive                     Direction A: Captology should focus on interactive
   technologiesare similar to the ethical implications for other                technologies that change behaviors.
   persuaders society (e.g., salespeople, coaches,    counselors,               As statedearlier, my definition of persuasionis “an attemptto
   religious leaders,etc.) [15]. Becausevalues vary widely, no                  shape,reinforce, or changebehaviors, feelings, or thoughts
   single ethical system or set of guidelines will serve in all                 about an issue, object, or action.” Although this is a good
   cases,so the key for those designing persuasiveinteractive                                                                               and
                                                                                definition, it is too broad to be useful in focusing research
   technologies is gaining a sensitivity to the ethical                         designin the early stagesof captology.Therefore,at this point
   implications of their efforts. High-tech persuaders  would do                people involved in captology would do well to pursue
   well to basetheir designson a defensibleethical standardAt                   behavior      change as the test and metric for persuasive
   the very least,a few core valuesshould apply to all persuasive               computers.
   computing designs, such as avoiding deception, respecting
                                                                                Behavior changeis a more compelling metric than attitude
   individual privacy, andenhancingpersonalfreedom.
                                                                                change for at least three reasons: (1) behavior change is
                                                                                thought to be more difficult to achieve than attitude change
   Ethics     of studying            persuasive      computers                  [15, 361, (2) behavior change is more useful to people
   The ethical implications for those who study persuasive                      concerned   with real-world outcomes[9, 31, 361, and (3) one
   computersare somewhatdifferent from those for designers-I                    can measure    behavior changewithout relying on self-reports
   proposethat those who study persuasivetechnologieshave a                     [9] (attitudechangemeasures     hinge on self-reports).
   responsibility to play a watchdog role for the HCI
   community, in particular, and for technologyusers,in general.                Direction B: Captoiogy should follow the well-established CHI
   Ethical actionsfor those who study persuasivecomputersfall                   tradition of adopting and adapfing theories and frameworks
   into four categories:                                                        from other fields.
   1. Identify     artifacts    and techniques. By using the                    While captologyhasthe potential to generatenew theoriesand
   various frameworkssuggested this paper, a researcher
                                   in                      can                                               to             in
                                                                                Cameworks-as demonstrated some degree this paper-
   better identify persuasive technologies, as well as the                      those of us who study computersas persuasivetechnologies
   persuasive  strategies technologyuses.
                          a                                                     would do well to find extant persuasion theories and
   2. Examine        effectiveness    and effects. Researchers
                                                                                frameworksandinvestigatehow they apply to captology. For
                                                                                example,Aristotle certainly did not have computersin mind
                        the             (the
   should also assess effectiveness intendedimpact) and
                                                     This is trial version
                                     7 ‘->Te   ..    ;:   .          _.         .~    *._.   .I ..i:.,     ‘:‘,,‘,..   _,   .e   *.   .‘.a;   ,.   ~I._   ’   ;    -. -.-,&     1
         _   --   -
CHI 98 e 18-23        APRIL   1998

 when he wrote about the art of persuasion, but the ancient               studies;as a result, no useful theory comesfrom this type of

 field of rhetoric can apply to captology in interesting ways.            research [21].
 The field of psychology-both cognitive and social-has a                  In order to avoid this well-traveled dead-endroad, those
 tradition of examining different types of persuasion and                 involved in captology research  should comparethe persuasive
 influence. The theories and methods from psychology transfer             effects of one aspect of computing technology with other
 well to captology, enriching the field. In addition, the field of        aspectsof computing technology.
 communication has a history of examining the persuasive
 effects of media and other types of message sources.                     Direction F: Capto/ogy should focus on both “what is” and
 Specifically, the applied domain of public information                   “what could be.”
 campaignshasa set of theories and practicesthat give insight             Captology should focus both on existing persuasive
 into the study of persuasivecomputers.                                   technologiesand on potential persuasivetechnologies.A fairly
                                                                          good number of persuasive technologies already exist.
 Direction C: Captology should examine and inform the design              Researchinto the impact, uses, and implications of these
 of interactive technologies that are specialized, distributed, or        technologiescan yield valuable insight for the field.
                                                                          But captology also has a strong generative component.
 While tbe typical computerof today lives on top of a desk and                     this
                                                                          Because field provides a new lens for viewing computing
 rquires a keyboard and a monitor, people involved in                     technology, captology generates insights about novel
 captology would do well to focus mainly on technologiesthat              possibilities for computers, as well as new ways to bring
 are specialized, distributed,     or embedded [for more on               aboutattitude andbehavior changes.
 the conceptof ubiquitous computing, seereferences and 341.
 From my vantage point, the most interesting interactive                  Direction G: Captology should be pursued with ethical issues
 technologiestoday seem to fall in at least one of these three            and implications in mind.
 categories. And it seems that most persuasivetechnologies of             The study of computers as persuasive technologies raises
 the future will not be associatedwith desktopcomputers;they              important ethical questions. This is understandableand
 will be specialized,distributed, or embeddedIf this is trne,             inescapable,given the fact that persuasionhas long been an
 then it would be a relatively poor use of time to examine                areafor ethical debate,and that computertechnologyhas raised
 existing desktopapplications or to designpersuasiveprograms              recent ethical questions. As with any other means of
 solely for desktop computers. Persuasive situations occnr                persuasion, one could compromise values and ethics
 most frequently in the context of normal life activities-not             intentionally or unintentionally. Therefore, those who study
 when peopleare seatedat a desktopcomputer-                               persuasivetechnologiesshould have a sound understanding       of
                                                                          the ethical implications of this field. A subfield on the ethics
 Direction D: Captofogy shou/d focus on endogenously                      of persuasivecomputing is a worthy and important endeavor.
 persuasive technologies.
 I’ve proposed three types of persuasive technologies:                    SUMMARY
 endogenously   persuasive(dose designedwith persuasivegoals              This paper has defined captology and has articulated five
 in mind), exogenously persuasive (those used to persuade                 different perspectivesfor studying computers as persuasive
 anotherperson), and autogenously persuasive(those used to                technologies. This paper also proposedseven directions for
 persuade oneself).         Understanding endogenously                    future research. While these perspectives and research
 persuasive      technologies     seems more essential to                 directions may be supersededby better ideas and new
 understandingcaptolo,v than the other two types. According               perspectives, this paper lays the groundwork for taking
 to my definition, endogenouslypersuasivetechnologies were                significant stepsforward in understandingthis new field.
 created with an intent to changeattitudes or behaviors. As a
 result, the strategiesand techniquesto persuadeare embedded
 in the technology itself. In contrast, the other two types of
 persuasive technologies-exogenous and autogenous-rely                    ACKNOWLEDGMENTS
 heavily on externalfactorsfor their persuasivepower.                     Cliff Nass,Byron Reeves,Terry Winograd, and Phil Zimbardo
                                                                          have beeninstrumental in my work on persuasivecomputing.
 Direction E: CaptoIogy can learn from other media but should
 steer clear of comparing computerpersuasion with persuasion
 from other media.                                                        REFERENCES
 Although captology has much to gain from other media
 research(suchas effectsof video, print, radio, etc.), comparing          Ul         Atkins, C., & Wallace, L. Mass Communication and
 computer persuasion with persuasion from other sources                              Public Health. Newbury Park, CA: Sage,1990.
 seemsto be a deadend. Many studies have attemptedto do
 this, and virtually all of them contain serious weaknesses [for          PI         Bandura,A. Social Foundations of Thought and Action:
                                                                                     A social cognitive theov. Englewoods Cliffs, NJ:
 a longer discussionof this claim, see 141.
                                                                                     Prentice Hall, 1986.
 Although a researchercan clearly determine that computer
                                                                          131        Buxton, W. “Living in Augmented Reality: Ubiquitous
 pro,= X is more persuasivethan video Y or pamphlet Z,
                                                                                     media and reactive environments.” Web URL:
 these results hold only for artifacts X, Y, and Z-not for
                                                    This is trial version
 comparing computers,videos, and pamphlets in general.One
 problem is that too many variables are at play in cross-media


   r41    Fogg, B.J. & Nass, C-1. “How Users Reciprocateto          WI
                                                                                                 CHI   98   l   18-23

                                                                         Nass, C. I. and Mason, L. “On the Study of
                                                                                                                        APRIL   1998

          Computers:An experimentthat demonstrates       behavior        TechnologyandTask:A variable-based  approach.”In J.
          change.”In Extended Abstract-s of CH1’97, ACM Press,           Fulk & C. Steineld (eds.), Organizations and
          p. 331-332.                                                    Communication Technology, 46-67. Newbury Park,
          Fogg, B.J. & Nass, C-1. “Silicon Sycophants: The               CA: Sage,1990.
          effectsof computersthat flatter.” International Journal   P21 Patton, P. “User interfacedesign: Making metaphors.”
          of Human-Computer Studies (1997), 46,551-561.                  I.D. Magazine, April, 1993, 62-65.
   161    Fogg, B.J. “Captology: The Study of Computers as          1231 Rawls, J. A Theory of Justice. Cambridge: Harvard
          PersuasiveTechnologies.” In Extended Abstracts of              University Press,1971.
          CHI’97, ACM Press,p. 129.
                                                                    [241 Reardon, K. Persuasion in Practice. Newbury Park,
   171    Friedman, B. Human Values and the Design of                    CA: Sage,1991.
          Computer Technology. New York: Cambridge                       Reeves, B. & Nass, C. The Media Equation: How
          University, 1997.                                                people treat computers, television, and new media like
   PI     Friedman, B., and Kahn, P. “Human Agency and                     real people and places. New York: Cambridge
          Responsible Computing: Implications for Computer                 University Press,1996.
          SystemsDesign.” Journal of Systems Sofnuare, 17:7-               Rice, R. and Atkin, C. Public Communication
          14, 1992.                                                        Campaigns. Newbury Park, CA: Sage,1989.
   PI     Graeff, J., Elder, J., andBooth, E. Communication for
                                                                    VI     Salmon, C. Information Campaigns : Balancing social
          Health and Behavior Change. San Francisco, CA:                   values and social change. Newbury Park, CA: Sage
          Jossey-Bass,   1993.                                             Publications, 1989.
   [lOI   Honeycutt, L. Resources on Computer-Supported             PSI                   B.
                                                                           Schneiderman, Designing the User Intelface. New
          Cooperative Work http://www.dcr.rpi.edulcscw.html                York: Addison Wesley, 1987.
   Ull    Huddy, C., Herbert, J., Hyner, G., & Johnson, R.                 Sheridan, T. Telerobotics, Automation, and Human
          “Facilitating changes in exercise behavior.”                     Supervisory Control. Cambridge, MA: MIT Press,
          Psychological Reports, 1995,76(3), 867-875.                      1992.
   WI     Kay, A. “Computer software.” Scientific American,         1301   Steuer, J. “Defining virtual reality: Dimensions
          251, (1984), 53-59.                                              determiningtelepresence.” Joumal of Communication,
   u31    Kozma, R. ‘Zeaming with media.” Review of                        1992,42(4).
          Educational Research, 61, (1991), 179-211.                       Street, R., Gold, W., and Manning, T. Health
   1141   Kuomi, J. ‘Media comparison and deployment: A                    Promotion ond Interactive Technology. Mahwah, NJ:
          practitioner’s view.” British Journal of l2iucational            LawrenceErlbaum, 1997.
          Technology, 1994,25(l), 41-57.                                   Turkle, S. ‘Who am we?” Wired Magazine, 1996,
   1151   Larson, C. Persuasion Reception and responsibility,              January,148-202.
          7th ed. Belmont, CA: WadsworthPublishing, 1995.           I331   Verplank, B., Fulton, J., Black, A., & Moggridge, B.
   WI     Laurel, B. The Art of Human-Computer Inter&e                     Observation and Invention: Use of Scenarios in
          Design. Reading,MA: Addison-Wesley,1990.                         Interaction Design. Tutorial at lNTERCHI’93,
          Lieberman, D. ‘Interactive Video Games for Health              Amsterdam,1993.
          Promotion.” In Health Promotion and Interactive           [341 Weiser, M. “The computer for the 21st century.”
          Technology, eds, R. Street, W. Gold, & T. Manning.             Scientific American, 1991,265(3), 94-104.
          Mahwah,NJ: LawrenceEarlbaum, 1997.                        I351 Winograd, T. ‘“The design of interaction.” In P.
   WI     Marcus,A. “Human communication issues in advances              Denning & R. Metcalfe (Eds.). Beyond Calculation:
          UIs.” Communications of the ACM, 36(4), 1993, lOl-             The nextfifty years of computing (pp. 149-161) New
          109.                                                           York: Springer-Verlag,1997.
   WI     Mutschler, E. “Computer assisteddecision making.”         1361 Zimbardo, P. and Leippe, M. The Psychology of
          Computers in Human Services. 19906(4), 231-250.                Attitude Change and Social Influence. New York:
          Nass, C. I., Fogg, B. J., Moon, Y. “Can computersbe            McGraw Hill, 1991.
          teammates?”    intemational Joumal of Human-Computer
          Studies, 45, 1996,669-678.

                                       This is trial version

To top