CHI '95 ProceedingsTopIndexes
PapersTOC

Ethics, Lies and Videotape...

Wendy E. Mackay

Rank Xerox Research Centre, Cambridge (EuroPARC)
61 Regent Street
Cambridge, UK CB2 1AB
mackay.chi@xerox.com

© ACM

Abstract

Videotape has become one of the CHI community's mostuseful technologies: it allows us to analyze users' interactions with computers,prototype new interfaces, and present the results of our research andtechnical innovations to others. But video is a double-edged sword. It isoften misused, however unintentionally. How can we use it well, without compromising our integrity? This paper presents actual examples of questionable videotaping practices. Next, it explains why we cannot simply borrow ethical guidelines from otherprofessions. It concludes with a proposal for developing usable ethical guidelines for the capture, analysis andpresentation of video.

Keywords:

HCI professional issues, video editing,ethics, social computing.

Introduction

The lights dim in the plenary talk at CHI'95. You settle back in your seatto hear from one of the early innovators in HCI - in fact, your formerthesis advisor from a decade ago. As expected, he is an entertainingspeaker. He quickly has the audience laughing as he shows videos of early interfaces and very perplexed users. Suddenly, you're not laughing. You see a familiar face projected on the 40 footscreen: it's you, ten years ago. You watch in horror as the 2500 membersof the audience, now your peers and colleagues, laugh at your 'inept' useof the technology. Could such a thing happen? It already has. What was the appropriate thingto do? Should the speaker have tried to discover if she were in theaudience? Would 'informed consent' given ten years ago have been adequate?What were her rights? What was the audience's responsibility? These are not easy questions and I won't presume to provide definitiveanswers. However, I think such examples can raise awareness of the issuesfacing the CHI community, as we increase our use of video for a wide rangeof activities. Sometimes, simply being sensitive to the problem is sufficient; other times, there is no clearcourse of action. In either case, I contend that we are obligated as aprofession to try to deal with these issues as effectively as possible.

As a community, we must educate ourselves aboutpotential misuse and encourage responsible behavior. We must alsounderstand who we are trying to protect and the trade- offs in protecting one groupversus another. We need comprehensive guidelines to help members of the HCIcommunity make ethical decisions. The next section challenges the perception that videocan be treated as an objective record of events and then presents examplesof questionable videotaping practices. I also discuss why the advent ofdigital video increases the potential for misuse. The subsequent section frames the discussion within amore general ethical framework. I briefly review the perspectives of otherprofessional groups, particularly with respect to their use of video. Thelast section presents preliminary suggestions for handling video and proposes a strategy for developing moredetailed guidelines for the HCI community.

VIDEO: OBJECTIVE ORSUBJECTIVE?

Video is a powerful medium: it can makea point or convince people in ways that other media cannot. Video captures aspects of human behavior, such as gaze and body language, that are not available in any other form. Somehow, video seems "real". Yet,perhaps it is too powerful. Just as statements taken out of context can bevery damaging, so can video clips misconstrue events or violate the privacy of the subjects involved.

Researchers often treat videotaped records of human behavior as objective scientific data: they can be viewed repeatedly,individual events can be counted and findings can be verified independently by other researchers. Unfortunately, the appearance of objectivity is just that: an appearance. Someone must choose a locationand field of view for the camera, which must include some and exclude otherinformation. The choice of when to press the "record" button also includes and excludes information. More subtly, the context shared by the participants of thevideotape may be difficult or impossible to capture and present to subsequent viewers.

The shared context can occur at various levels. Forexample, Clark & Schaefer (1989) examined conversations between people. Ifone person is explaining something, she looks to the other person forsigns, such as a nod or "uh huh" that he has understood sufficiently well for her tocontinue. She may not speak clearly but will continue if she is convinced that he is followingher. Is she misspeaks, she may see him look puzzled and then smile,indicating that he has understood and she should continue. A camera shot ofher face as she speaks will capture the exact words she spoke but not the shared understanding that evolved. Thevideo records only the fact that she misspoke. Later, it could be usedto "prove" that it was what she "really" meant.

Another problem arises when video captures conversations between peoplewith shared prior experience, who speak in short-hand. In a live setting,an observer might be puzzled by what is meant or ask for clarification.With a video record, the same observer could view it repeatedly, develop a theory about the meaning and becomeconvinced she understands, even if the participants meant somethingelse. People are used to being able to speak informally in daily conversation.Since both speakers and listeners know their memories can be unreliable,misunderstandings are usually cleared up through further discussion. When casual conversations are recorded, the ways of resolving misunderstandings changes. Suddenly, the speaker canno longer say "I didn't say that"; the videotapedrecord becomes an independent arbiter of what was said. But what was said is not the same as what was meant. Since peoplecan change their minds over the course of a conversation, statements thatseem to establish what the speaker 'really' meant distort the ongoing process of conversation.

Most people (except for politicians ) feel uncomfortable being recorded and change their usual behavior; theyare not used to speaking "for the record". If electronic mail is notorious for generating misunderstandings due to informal writing, recorded casual speech is worse. Even speaking carefully can be dangerous, since viewers may interpret it as evidence of 'something to hide'. Broadcast media are thus subject to greater restrictions than print media. For example, "Recognizingthe particular power of radio and television to influence public opinion, federal legislation was passedlimiting the involvement of broadcasters in political camps." (Hall,1978)

Recording video is only part of the problem. The audience and context in which the video is presented may alsoaffect what is understood. For example, imagine recording a researcher'sdiscussion of a new software interface that 'increases productivity'. Thisvideo, shown to employees who interpret "productivity" as a euphemism for layoffs, suddenly has a very different impact. Theinfamous "sound bite", in which a short clip is selected to represent a longer event, maydistort the original message or make rare events appear representative. "TV news often avoids coverage of the story that doesn't have anything visual and too often makes editorial decisions based on the availability of pictures rather than true news value." (Hall, 1978, p.17) These examples demonstrate the importance of context and how easily videocan be misinterpreted, intentionally or not. Unfortunately, even people whorecognize that a videotape is not an objective record find it easy to slipinto thinking that it is somehow real. Video is powerful; care is required both in its production and its interpretation. The use of video raises ethical questions: we can look tothe literature in ethical theory for help addressing them.

ETHICAL THEORY

The ethical literature is vast, with philosophical discussions dating backto Plato and Aristotle. According to Forester and Morrison (1990), mostcurrent professional ethical codes are influenced by three more modernperspectives: ethical relativism (Spinoza), consequentialism or utilitarianism (J.S. Mill) and deontologism(Kant). The latter two are most relevant for computer professionals:"Consequentialismsays simply that an action is right or wrong depending upon itsconsequences, such as its effects on society. [...] By contrast,deontologism says that an action is right or wrong in itself. Deontologistsstress the intrinsic character of an act and disregard motives or consequences." (Forester and Morrison, 1990, pp.16-17) Older, more established professions, such as medicineand law, provide codes of ethical practice for their members. Their goalsare to establish their status as a profession, to regulate their membership and convince the public that they deserve to beself-regulating (Frankel, 1989). Some, such as Ladd (1980), dismiss thenotion of organized professional ethics as having few benefits and realpotential for harm, while others, such as Bagley (1977), argue that "a written code is a necessity".Luegenbiehl (1992) argues that "Codes of ethics need be neither authoritarian nor designed for theenhancement of a profession. Instead, they should help the professionalseeking to engage in ethical practice". Computer science is a relatively new field but already has a largeliterature on ethics and computing. (See recent books by Forester andMorrison (1990), Johnson (1994) and Dunlop and Kling (1991).) Martin andMartin (1994) compare four codes of ethics: ACM (1992), IEEE (1992), Data Processing Managers Association (DPMA, 1989)and the Institute for Certification of Computer Professionals (ICCP, 1989).The four codes are similar to each other and to other professional codesbecause they take ageneric approach to ethics. Privacy and confidentiality of data were seenas the only elements that "reflect the unique ethical problemsraised by computer technology" (Martin and Martin, 1994). Since video involves both privacy andconfidentiality issues, ethical guidelines for HCI must go beyond generalethical codes.

The ACM Code of Ethics and Professional Conduct,revised in 1992, is generally considered to be the most complete. Andersonet al. (1992) state that the new ACM code "recognized the difficulty that ACM and other societies have in implementing an ethics reviewsystem and came to realize that self-regulation depends mostly on theconsensus and commitment of its members to ethical behavior". Like Luegenbiehl (1992), they argue that the most important function of a code of ethics is its role as an aid to individualdecision-making. They illustrate ethical issues with nine cases that callfor individuals to make ethical decisions. Each case has an individualscenario illustrating a typical decision point that relates to sections of the code. Bok (1982) reported that over 12,000 distinct ethics courses, includinglaw, medicine, business, engineering, liberal arts, research sciences,religion and philosophy, were taught in American academic institutions.Discussing case studies in the class room has been shown to be an effective teaching approach (Dunfee, 1986) and theSIGCAS newsletter regularly presents suchethical case studies fordiscussion (e.g., Gotterbarn, 1993). Rather than argue about the merits of different ethical philosophies, I have chosen to follow thisstrategy, presenting scenarios based on real events and proposingguidelines related to the capture, production and presentation of video.

QUESTIONABLE USES OF VIDEO

The following examples of questionableuses of video are based on actual incidents. However, some of the details have been changed to disguise the participants or setting.

Candid Camera? Linda is preparing her CHI'95 presentation and wants to give an entertaining talk. She looks through her videotapes of user sessions and finds several funny clipsof users doing unexpected things. At the talk, she makes a joke and showsthe clip; the audience laughs. Is Linda guilty of perpetuating a "candid camera" approach, in which research videos become transformed into a form ofentertainment at the expense of users? Is this an appropriate activity forprofessionals who purport to support users? On the other hand, does this mean that we can't have entertaining CHI talks or videos?

Lack of permission? Jane is a trained anthropologist who has just conducted a study of workpractices within a corporation. She and her colleagues have videotaped anumber of meetings in which sensitive issues, such as determining whoshould be laid off, have been discussed.The participants are very sensitive about being videotaped and haverequested that the videotape not be shown to anyone else in the company.Later that year, Jane presents her work to at a workshop at a CHIconference and includes several clips of video taken from her research. Is this a violation of her agreement with theparticipants in her study? Is there a way in which she can disguise thevideo to prevent any possible feedback from the research audience to thecompany?

Is the reviewer responsible? Ralph is reviewing presentations for a workshop he is running. Several ofthe participants propose to show video of users involved in their work. Hedecides that it is the responsibility of the authors to obtain theappropriate permissions and does not ask whether the authors have permission to present the tapes inthis forum. What is the reviewer's role? Should he remind theauthors of their obligations? Should he go further and request evidence ofhaving obtained appropriate permissions? Under what circumstances should hereject a submission?

Wrong audience? Fred is developing a technique forcombining real data with video simulations to provide training for pilots.He takes data from the flight recorders of planes that have crashed andrecreates the situation, including external weather conditions and instrumentreadings. He plays one of his recreated videos to human factors colleagues,who suddenly find themselves listening to the voice of a real pilot saying:"Oh my God!" followed by a scream and a crash. The audience is stunned. Suddenly the very personalexperience of another human being's death was being presented to them,without warning, as a part of a training exercise.

Was it appropriate to show a sensitive video designedfor one audience to another? Was this a violation of the dead pilot'sprivacy? Could he have presented his work to this audience without usingthe real tape?

Undue influence? Harry conducts usability studies of new software products for hiscorporation. He videotapes each usability session and carefully analyzeswhat causes the user's problems and where they make errors. He thendiscusses the issues with the software developers. Harry is particularly annoyed by one feature and wants to convince thesoftware developer that it should be changed in a particular way. He shows a video clip of oneof the users struggling with the feature as proof that his way is better.He does not show other clips in which users do not experience problems withthe feature. Is Harry taking advantage of people's willingness to think that video is anobjective record in order to win an argument? Could Harry provide a morebalanced view by presenting an overview of the relevant anecdotes? Whatwould such an overview consist of?

Inappropriate special effects? John is preparing a video of his newsoftware system for the CHI'95 conference. He carefully records whathappens on the screen and then edits out a number of "boring" sections in which the system responds especially slowly. He adds a cut toa separate system, which will eventually be integrated with his, to showwhat would happen if they were connected.

Under what circumstances is it reasonable to make asystem appear faster or more complete than it is? Would a disclaimer,describing the level of editing, be sufficient?

Inappropriate reuse? Mary is the product manager in charge of a new product being exhibited forthe first time at CHI'95. She is proud of their usability lab and showsvideotapes of some of the user studies to illustrate how well the interfaceworks. When asked if she had obtained permission from the subjects of the video, she is surprised and saysit had not occurred to her to do so. She believes she is safe, legally,since the people in the tape were company employees.

Even if she is not legally liable, does Mary have a responsibility to ask permission from the subjects? When is it appropriate to ask permission? Prior to recording, after the subject has seen the video, or just before each event in which the video will be shown. Is it possible for the subject to really understand what the implications of giving permission are?

Recording without permission? The XYZ research laboratory allowspeople in the lab to communicate with each other via live video connections. Privacy issues have been carefully considered and there are a variety of ways for peopleto select how others may connect to their cameras. A separate program takes snapshots every few minutes from the media space and displays them in awindow. One day, one of the participants in the media space walks into a room where a group of her colleagues is laughing at something. She discovers it's a picture of her,with someone giving her a kiss on the cheek (actually, her husband). Sinceit is impossible to see who the person is, the group laughingly teases her about who it might be.

What is the difference between a temporary record, inwhich a recently-shot image is displayed, and a more permanent record? Is it acceptable to select segments from an on-going stream of activity and highlight them?

Computing on video All the previous examples have actuallyoccurred, based on today's technology. We face a potentially much biggerproblem with the advent of digital video. At SIGGRAPH '93, a panel of special effects experts showed a"behind the scenes" look at Jurassic Park, in which a stunt woman's image is changed to become that of the main actress. We fully expect special effects in science fiction movies and are amazed by the skill at which dinosaurs can be madeto look real. What is less obvious is that special effects are used in most Hollywood movies to create images of reality. These techniques can be used to distort whatwe see.

Employers already monitor workers through computers. Pillar (1993) surveyed over 300 CEOs and MIS directors and found that 22% searched electronic mail, voice mail,computer files and other networking communications of their employees. Lyon(1994) discusses the role of electronic surveillance in society. Video isincreasingly part of that electronic surveillance. For example, Great Britain has a new system that automaticallyreads the number plates (license plates) of a speeding car and displays the number, together with the excess speed, on a roadside display. The aim at present is to shame the offender, but the next step may be to link the system to a police database. In the past, people had to watch video from electronic surveillance cameras. Now, computers can watch for us.

The above list is not exhaustive, but illustrates problems of varying levels of severity. In most of these examples, the individuals are well-intentioned. In fact, some members of the HCI community will find nothing wrong with some of these scenarios. But this makes the issue problematic: we need to raise the level of awareness and try to establish guidelines that we can agree upon.

GUIDELINES FROM OTHER PROFESSIONS

Since Human-Computer Interaction is a new field, we should learn fromother, more established professions. Some research disciplines,particularly the medical and social sciences, have well-establishedguidelines for using human subjects and include the use of videotaped records in this context. Other disciplines, such ascomputer science, have no history of using video (or human subjects),leaving HCI members from those fields without any guidance. Unfortunately,even those disciplines that do have guidelines for video do not provide sufficient guidance for thediversity of uses of video found in the HCI community. This section brieflysummarizes the ethical or legal perspectives of various professions.

Medicine Physicians have a long history ofdealing with ethical issues. The Hippocratic oath urges physicians to"do no harm", i.e. to protect the patient. Key issues include who should choose apatient's treatment planand how can patients without medical training evaluate risks or giveinformed consent about procedures. Doctors must present the options andsupply all "material" information to the patient, but not necessarilyprovide full disclosure. Macklin (1987, p.45) describes the evolution of biomedical codes from the professionalcommunity standard, which asks "what reasonable medical practitioners insimilar situations would tell their patients" to the current reasonablepatient standard: "what the reasonable patient would want to know before giving consent to a recommended therapy."Studies show that poor communication and lack of information make patientsmore likely to refuse a particular treatment. This standard has helpeddoctors develop better relationships with their patients, with the accompanying danger that better relationshipsmake it easier to obtain consent. Shannon (1976) and Beauchamp & Childress(1983) provide different views on biomedical ethics. Collste (1992)explores the question of whether computers, particularly expert systems, cause new moral problems.

Social Sciences Experimental Psychologists who perform experiments with people are expectedto follow guidelines established by the American Psychological Association(1991) or the relevant organization in other countries. Individualuniversities and organizations often publish guidelines, e.g., Queen's University (1989) or UCLA (1987).Most universities also have a committee that reviews research proposals andapproves the procedures, e.g., the Massachusetts Institute of Technology Human Subjects Review Committee.

Subjects in Psychology experiments must sign a consentform that describes how any data collected about the subject will be used.After the experiment is completed, the experimenter is expected to"debrief" the subject and explain what occurred. Most guidelines are designed toprotect the subject from harm. The APA guidelines were influenced by afamous set of experiments by Milgram (1965). Subjects were told toadminister electric shocks to people (actually confederates of the experimenter) if they missedquestions on a learning test. Milgram found that subjects followed theseorders, even to the extent of believing they had killed the personreceiving the shocks. Understandably, the subjects were traumatized by this experience.

Anthropologists and Sociologists work with people infield rather than laboratory settings. Videotape is increasingly used torecord people's activities in the context of their daily lives. Bothprofessions have also established ethical guidelines for the protection of their subjects.Critical issues include the problems of how to handle data collected in thefield and how to handle naive informants who may not be able to give trueinformed consent.

Journalism Hulteng (1985) describes the chief function of journalism as "thecommunication to the public of a reasonably accurate and complete pictureof the world around us [...] The central ruling ethic of journalism [is] toreport the news of the world dependablyand honestly." (pp. 170-171) Broadcast journalists are thus ethically beholden to their audiences: they "protect" their viewers by presenting an"objective" account of an event. It is ethical to show a person negatively,as long as it is a "truthful" view. However, Hall (1978) explains that the FCC requires journalists to "contact the person attacked, provide a transcript of the charge and allow equal time for a response." Ordinary people (i.e., not celebrities) may not have their images broadcast without permission, unless the event is 'news' that occurred within the past 24 hours. Hall discusses journalist's rights and responsibilities, from the FairnessDoctrine, which covers libel, slander and invasion of privacy to the Shieldand Sunshine laws, which enable journalists to protect their sources. Kronewetter (1988), as well as Hulteng and Hall, discuss journalism ethics and Malcolm (1990) and Alley(1977) provide exposÚs of ethical violations.

Documentary film-makers do not believe in a single,objective point of view. Their goal is to present a fair perspective, from aparticular point of view, through selective shooting and editing.Participants in their films should feel they have been presented fairly,if not always positively.

Marketing Firms Marketing firms videotape "focus groups" to get customer reactions to new and existing products. Their loyalty isto producer of the products they examine. They must protect their clients,not only from potential lawsuits but also from information leaks tocompetitors.

Law and Accounting Firms "Lawyer-client privilege" and"accountant-client privilege" (Causey, 1988) enable clients to speak inconfidence to these professionals, another case of protecting the client,both legally and through ethical codes.

Publishers Publishers must obtain copyright permission from the person who created the videotape before theycan distribute it. They are legally responsible for protecting the producer (or copyright holder) of the videotape. Samuelson (1994) discusses legal precedents for the fair use of copyrighted material, including video, e.g.,the ability of consumers to videotape broadcast television programs for home use.

Software Developers and Other Corporations Corporations use video for a variety of purposes, from usability studies to product marketing. Getting permission protects the corporation from lawsuits. Hollywood's Universal Studios obtains global permission fromtheir visitors: a sign informs them that, by entering the park, they have given tacit permission to be videotape dand their images may be used for commercial purposes. People who object aredirected to a guest relations office.

Who are you trying to protect? Trying to understand the goals of each of these professional guidelinesreveals a fundamental problem: each is concerned with protecting someone,but they are all different types of people. Some try to protect the personbeing videotaped. Others try to create an objective view for the benefit of an audience. Some must protect theconfidentialityof their clients, while others want to protect the producer of thevideotape. The HCI community includes people concerned with each of thesesituations; our ethical guidelines must somehow address them all.

PRELIMINARY GUIDELINES

Who should the HCI community listen to when developing ethical guidelinesfor video? We have a diverse (and growing) set of uses of video, both asdata about users and technology and as a presentation form for users,customers, management, fellow developers and the HCI research community. What perspective or perspectives should we consider? It is notenough to simply say we should "protect everyone"; we might end upavoiding video all together. We must consider the implications of avariety of uses of video and develop guidelines accordingly.

A good set of guidelines must cover everything from the initial videotapingto its final presentation and address, at least, the following questions:How do we obtain "informed consent"? How should recording of video be constrained? Are restrictions on the analyses performed necessary? Under what conditionsshould video be presented and to which audiences? Who are we trying toprotect? How can people protect themselves and what social structures areneeded to ensure that they can? What are the legal and cultural implications of videotaping in differentcountries? How do we avoid confusing ethics and good taste?

The suggestions presented below are offered as astarting point for discussion, rather than a definitive set of guide- lines.Theyare based on discussions with members of the HCI community and influencedby guidelines from other professions. I encourage people to try them andprovide feedback about what does and does not work.

For the purposes of clarity, the term producer isused to refer to any person who creates a videotape, including academicresearchers, usability specialists and software developers. The term user refers to any person in the videotape, including participants inlaboratory studies or people being videotaped in the course of their dailyactivities.

A. Prior to Recording

  1. Establish what constitutesinformed consent Prior to recording, obtain informedconsent : make sure the user understands the implications of being videotaped. Theproducer must define what constitutes informed con- sent. This may bedifficult, as in the introductory example.
  2. Inform people of thepresence of live cameras If a camera is left on, e.g., in a mediaspace or to record an event, let people know when they are on camera andgive them the opportunity to avoid being in the camera's view. A sign should state whether or not the video is being recorded. Forexample, EuroPARC's media space uses a camera in the commons area. Amannequin holds the camera and a sign to let visitors know they are oncamera.
  3. Ask for permission beforevideotaping Tell users that a videotape record willbe made and give them the opportunity to speak off the record or stop therecording altogether. Consider if the user feels social pressure to agreeand make it clear that saying no is legitimate. Avoiding social consequences may bedifficult, e.g., when a meeting is taped and only one person objects.
  4. Explain the purpose of the video Tell users the expected purpose and other potential uses of the video. For example, videotapes from usability studies are sometimes re-used for advertising. Tell users whether separate video clips or the entire sessioncould be used.
  5. Explain who will have accessto the video Tell users if anyone other than theproducer will view the video. Users may not mind a researcher seeing a tape, but may feel uncomfortable if it is shown to colleagues, managers or general audiences,e.g. at a CHI conference.
  6. Explain possible settingsfor showing the videotape Tell users where the videotape could be shown. For example, at CHI conferences, videotapes may be shown to large audiences during talks, in small videotape viewing rooms, or on the hotel cable TV. In some corporate settings, some video clips may be used for advertising.
  7. Explain possible consequences of showing the video Producers may find it difficult to adequately convey how a user might feel if the video were shown in a certain setting. For example, a video clipshown on a television monitor to colleagues might be acceptable, but highlyobjectionable when projected on a 40 foot screen to a large audience.
  8. Describe potential waysvideo might be disguised If the video will be used inunpredictable settings, describe how the user's image will be disguised,e.g., through blur- ring the user's face. Mantei's (1990) "Strauss Mouse" video is a clever example of avoiding potentially embarrassing use of research videos; she used actors' hands to demonstrate the ways executivesmisunderstood a 'simple' computer mouse.

B. After Recording

  1. Treat videotapes of users asconfidential Do not allow others to view videotapes casually and restrict access to them. This protects producers as well, e.g., if a manager decides to reuse video in ways that violate the original agreement between theuser and producer.
  2. Allow users to view videotapes Ideally, give the user the opportunity to view the completed video. If this is not possible, the producer should consider ways in which people can bedisguised. For example, some video editing systems can blur or distort aface.
  3. If use of the videotapechanges, obtain permission again Asking permission is not a simple matter. Permission can be given before recording or after the user has beentaped, or after the user has seen the tape, or just prior to an event in which it will be shown. The user can give blanket approval orapprove individual events.
Give users sufficient information to make an informed choice and let themchange their minds. For example, in the CHI'89 Kiosk (Soloman, 1990),users who contributed their images for the conference were again asked fortheir permission when the database was printed on a CD-ROM.

C. Editing Video

  1. Avoid misrepresenting data Producers are responsible for editingvideos so as not to imply that particular events are representative if they are not. If video is presented as data,distinguish between anecdotal and representative clips of "typical"events.
  2. Distinguish betweenenvisionments, working prototypes and finished products Clearly label presentations of technology as envisionments, workingprototypes or finished products. Envisionments propose or illustrate ideasthat have not been fully implemented. Working prototypes have beenimplemented and should not resort to tricks to make them look more complete. Products are completed commercial systems and must avoid misrepresentingtheir performance or features. For example, Wellner's (1992) videotapeincludes clearly labelled envisionments of future ideas contrasted withexamples of working software.
  3. Label any changes made toenhance technology Show the actual time it takes for aparticular operation or else clearly label cuts designed to improve thepacing of a video presentation. Do not simply cut out the slow sections tomake your system appear faster.

D. Presenting Video

  1. Protect users' privacy Hide individuals when possible. For example, shoot over the user's shoulderto see the screen, rather than the user's face. Obviously, this only worksif specific characteristics of the user, such as facial expressions, arenot an essential part of the record. Consider disguising the user's voice.
  2. Do not highlight clips thatmake users look foolish Do not show "funny" clips to make userslook foolish. This does not mean avoiding all amusing video clips; just besure that the joke is not at the user's expense.
  3. Educate the audience When giving a presentation, educate theaudience: rather than laughing at the user, explain how misconceptionsabout the technology can lead to breakdowns.
  4. Do not rely on the power ofvideo to make a weak point Be careful when showing video clips tosupport arguments in favor of particular technology changes. Some videoclips may magnify small problems or present a distorted picture.
  5. Summarize data fairly Clearly state the purpose of summariesof video data. Video data can be compressed in a variety of ways. Video clips canprovide a shortenedversion of what occurred in the session or can be used to "tell a story".If clips are presented in random order, they can be combined to show "typical" interactions, highlight unusual or importantevents, or present collections of interesting observations.

D. Distributing Video

  1. Do not use videos for purposes for which they were not intended Do not allow video of users to be usedfor purposes that they are not aware of, e.g. for an advertisement.

NEXT STEPS

ACM/SIGCHI has already begun to address a few of the issues relating tovideo. Every year, attendees ask to videotape CHI conference presentations,often for good reasons, such as non-native speakers who want a videobackup. The SIGCHI executive committeeis currently drafting a set of videotaping guidelines to try to balance theneeds of audience members with the rights of presenters. The vision.chi@xerox.com mailing list has been the forum for the discussion of various drafts and the final versionwill be published in the SIGCHI Bulletin.

Another policy statement on video appears in the CHICalls for Participation, e.g. from CHI'95: "Submission of video or pictures of identifiable people should be done withthe understanding that responsibility for the collection of appropriatepermissions rests with the submitter, not CHI'95." This gives submitters the unfortunate impression that this is solely alegal issue and that once permission has been obtained, the submitter andthe conference have no further responsibility in the matter.

The CHI community, given its mix of disciplines and variety of activities, has a unique perspective to offer on the issue of ethics and video. We should take advantage of CHI-sponsored conferences to raise awareness andgenerate discussions, e.g. Mackay (1989, 1990). We can establish an electronic discussion forum and consider collaborations with other organizations, such as SIGCAS (Computers and Society), CPSR (Computer Professionals for Social Responsibility) and theElectronic Frontier Foundation.

In the late 1980's, SIGCHI sponsored a task force that produced the influential ACM SIGCHI Curricula for Human-Computer Interaction (Hewett et al., 1992). Perhaps the time has come for a similar task force to develop an HCI code of ethics that builds upon the general ACM code and addresses issues unique to HCI, such as video.

CONCLUSIONS

This paper illustrates how easy it is,however inadvertently, to misuse video. Because videotape has become so prevalent in our profession, itis time for us as a community to become aware of the potential dangers and develop guidelines for ethical handling of video. These guidelines must go beyond legal requirements and provide protection for a variety of people involved in the HCI community.

HCI is not the only professional field that uses video. We can learn from other professional ethical codes. However, we cannotblindly adopt other ethical codes. Each profession is concerned with protecting someone: the person in the video, the audienceviewing the video, the client paying for the video or the producer of thevideo. Since the HCI community must address the needs of all of these people, we are uniquely positioned to create a broad- based set of guidelines that help us make informed, ethicaldecisions about our uses of video. If we are successful, guidelines may influence the wider set of organizations who are struggling with how to handle this powerful new medium.

ACKNOWLEDGEMENTS

The participants in the "Videoas a Research and Design Tool" Workshop and attendees of the CHI'90 Discussion Forum on Video Ethics provided many insights and examples. Austin Henderson pointed out the importance of educating audiences when presenting "funny" clips of users. Annette Adler, Sara Bly, and Marilyn Mantei contributedinteresting discussions and comments on earlier drafts. I am especiallygrateful to Michel Beaudouin-Lafon for reading earlier versions of thispaper and suggesting the title.

References

  1. ACM Code of Ethics andProfessional Conduct, (1993) Communications of the ACM, vol. 36:2, pp.100-105.
  2. Alley, R.S. (1977) Television: Ethics for Hire? NY: Abington Press.
  3. American Psychological society (1991)The importance of the citizen scientist in national science policy, APSObserver, vol. 4, pp. 10-23.
  4. Anderson, R.E., Johnson, D. G.,Gotterbarn, D. and Perrolle, J. (1993) Using the new ACM Code of Ethics indecision making. Communications of the ACM, vol. 36:2, pp. 98-105.
  5. Bagley, F.J. (1977) Ethics, unethicalengineers, and ASME. Mechanical Engineering. vol. 99, pp. 42-45.
  6. Beauchamp, T.L. and Childress, J.F. (1983) Principles of BiomedicalEthics, NY: Oxford University Press.
  7. Bok, D. (1982) Beyond the IvoryTower: Social Responsibility of the Modern University. Cambridge: HarvardUniversity Press.
  8. Clark, H.H. and Schaefer, E.F. (1989) Contributing to Discourse. Cognitive Science, vol. 13, pp. 259-294. Causey, P.Y. (1988) Duties and Liabilities of Public Accountants. NY:Dow-Jones Irwin.
  9. Collste, G. (1992) Expert Systems in Medicine and Moral Responsibility. Journal of Systems Software. Vol.17, pp. 15-22.
  10. DPMA (1989) DPMA Position Statement Handbook. 505 Busse Highway, Park Ridge, IL 60068.
  11. Dunfee (1986) Integrating ethics into the MBA core curriculum. The Wharton School of business. Reseachfunded by EXXON Educational Foundation.
  12. Dunlop, C. and Kling, R. (Eds.) (1991). Computerization and Controversy. NY: Academic Press.
  13. Forester, T. and Morrison, P. (1990) Computer Ethics: Cautionary Tales and Ethical Dilemmas in Computing.Cambridge: MIT Press.
  14. Frankel, M.S. (1989) Professional Codes: Why, how and with what impact? Journal of Business Ethics, Vol. 8:2-3, pp.109-116.
  15. Gotterbarn, D. (December 1993) CaseStudies: The Captain Knows Best. Computers and Society. Vol. 23:3.
  16. Hall, M.W. (1978) BroadcastJournalism: an Introduction to News Writing. NY: Comunication Arts Books,Hastings House.
  17. Hewett, T. (1992) ACM SIGCHICurricula for Human Computer Interaction, ACM, New York: NY.
  18. Hulteng, J.L. (1985) TheMessenger's Motives: Ethical Problems of the News Media. Englewood-Cliffs, NJ: Prentice-Hall .
  19. ICCP (1989) ICCP Code of Ethics. Your Guide to Certification as aComputer Professional. ICCP, 2200 E. Devon Ave., Suite 268, Des Plaines,IL 60018.
  20. IEEE (1979) IEEE Code of Ethics.IEEE, 345, East 47th St., NY, NY 10017.
  21. Johnson, D.G. (1994) Computer Ethics, Second Edition. EnglewoodCliffs,NJ: Prentice Hall. Kronewetter, (1988) JournalismEthics. NY: F. Watts.
  22. Ladd, J. (1980) The quest for a codeof professional ethics: intellectual and moral confusion. inR. Chalk,
  23. M.Frankel, & S. Chafer, Eds., AAAS Professional Ethics Project, American Association for the Advancement of Science, Washington, D.C., pp.154- 159.
  24. Luegenbiehl, H.C. (1992) Computerprofessionals: Moral autonomy and a code of ethics. Journal of SystemsSoftware. Vol. 17, pp. 61-68.
  25. Lyon, D. (1994) The ElectronicEye: The Rise of Surveillance Society. University of Minnesota Press.
  26. Mackay, W. (1989) Video as a Research and Design Tool. Special issue ofthe SIGCHI Bulletin, Vol. 21:2.
  27. Mackay W. (1990) Video Ethicsdiscussion forum, in Proceedings of CHI'90 Human Factors in ComputingSystems (Seattle, WA), ACM Press, pp. 381-386.
  28. Macklin, R. (1987) Mortal Choices:Bioethics in Today's World. NY: Pantheon Press.
  29. Mantei, M. (1990) The Strauss Mouse. SIGGRAPH Video Review.
  30. Martin, D.C. and Martin, D.H. (June 1994) Ethics Code Review.Computersand Society. ACM Press pp.21-25.
  31. Milgram, S. (1965) Some conditions of obedience and disobedience, HumanRelations, vol. 18, pp. 57-76.
  32. Pillar, C. (July 1993) Bosses withX-Ray eyes: Your Employer may be using computers to keep tabs on you. Macworld. pp. 118-123.
  33. Queen's University Department of Psychology (1989) EthicsClearance for Research on Human Subjects, Queen's University,Kingston, Ontario, Canada.
  34. UCLA Faculty Handbook, Supplement A: Policies and Procedures, (1987)University of California at Los Angeles Publications Office, Los Angeles,CA.
  35. Salomon, G. (1990) DesigningCasual-Use Hypertext: The CHI'89 InfoBooth. Proceedings of CHI'90Human Factors in Computing Systems (Seattle, WA) ACM Press. pp.451-458.
  36. Samuelson, P. (January 1994)Copyright's Fair Use Doctrine and Digital Data. Communicationsof the ACM, vol. 37:1, pp. 21-27.
  37. Shannon, T.A. (1976) Bioethics.NJ: Paulist Press. Wellner, P. (1992) The Digital Desk.SIGGRAPH Video Review.