ABSTRACT
The recent proliferation of video surveillance
cameras interconnected with high-speed computers and central databases
is moving us toward a high-speed "surveillance superhighway," as
cameras are used throughout entire cities to monitor citizens in
public areas. As businesses work alongside governments to build this
superhighway and expand it into private areas as well, there is a
growing need to develop methodologies of questioning these
practices. The goal of this paper is to stimulate inquiry into both
surveillance and the rhetoric used to justify its use. "Reflectionism"
is proposed as a new philosophical and tactical framework that takes
the Situationist tradition of appropriating the methodology of the
oppressor one step further by targeting that methodology directly
against the oppressor. The oppressor then becomes the audience of a
performance resulting from this new use of his or her own
methodology.
There is no place for the privacy factor when public safety is concerned.
-John Fitzgerald,
Supervisor, Transportation Operations,
U.S. Postal Service, New York
[1]
"A computer system being installed at welfare offices will compare each applicant's face to a database of thousands of other recipients' faces ... exposing fraud faster and more efficiently than other methods such as fingerprinting.... Viisage Technology, in Acton, bought the rights . . . and produced the fraud-detection system for the welfare department. Under its $112,500 state contract, Viisage will provide facial-recognition and fingerprinting services to welfare offices in Springfield and Lawrence as part of a six-month pilot program" [4].
Meanwhile, Privacy International is calling for a ban on Computerised Face Recognition [5] and ordinary citizens are arming themselves with ink pads and demanding that politicians and other officials submit to fingerprinting.
Although all of the
above uses of surveillance are associated with claims toward a better
future, an object of this paper is to ask the question "at what
price?" and to stimulate further inquiry into some of these
issues.
Embodied in the work presented in this paper is my
assertion of "acquisitional privacy," a concept that challenges the
right of organizations to capture/record images of an individual,
regardless of what promises are given regarding end use. Tacit in my
assertion is the notion of self-ownership [19]. (Some self-ownership pieces are
illustrated in Fig. 1)
A further goal of this paper is to call into question totalitarian visual surveillance. Totalitarian visual surveillance refers to a state of being in which individuals are "seen" by a remote and unobservable entity (human or machine) but do not "see" each other through the apparatus. (This situation calls to mind Jeremy Bentham's Panopticon [20], a structure he proposed for prisons, schools, workplaces and the like, where prisoners, students, employees, etc., could not see or interact with one another, but could be seen by a potential guard in a specially designed guard tower. The tower was designed so that individuals would not know whether the guard was watching them or even whether there actually was a guard in the tower).
One example of totalitarian video surveillance is found in department stores where extensive video surveillance is used, yet photography is prohibited. Of all forms of surveillance, totalitarian surveillance is particularly disturbing, as representatives of the video surveillance superhighway refuse to accept the accountability they demand-furthering us toward a Panopticon society in which we are treated more like prisoners than members of a community.
Important to the thesis of this paper are the following ways in which agents and representatives of the video surveillance superhighway defend their infrastructure: (1) Secrecy: the field is often not subject to open debate or peer-review; (2) Rhetoric: "public safety," "loss prevention," "For YOUR protection you are being videotaped"; (3) Constancy: department store clerks do not follow shoppers around with camcorders, but, rather, video surveillance is present in a "matter-of-fact" manner, as part of the architecture's prosthetic territory; (4) Higher and unquestionable authority: "I trust you and know you would never shoplift, but my manager installed the cameras," or "We trust you, but our insurance company requires the cameras"; (5) Criminalization of the critic: "Why are you so paranoid; you're not trying to steal something are you?"
One of my goals in applying Reflectionism to the surveillance problem is to allow representatives of the surveillance superhighway to see its absurdity and to confront the reality of what they are doing through direct action or through inaction (blind obedience to a higher and unquestionable authority).
In this sense, the video recording/transmission functionality of the apparatus appeared as incidental rather than intentional. When I wore the WearCam into an establishment, I did not give the impression that my purpose was to record video, partly because the apparatus was less visible than a traditional camera, but, more importantly, because the apparatus did not have the appearance of intentionality. In this way, the apparatus provided a mirrorlike symmetry between myself and those placing me under surveillance (e.g. shopkeepers' security guards): I was in a position to violate the privacy of representatives of an organization that was placing me under surveillance (e.g. representatives of a department store or the like) without violating their solitude (i.e. without an unusual form of interaction, as might be the case when using a hand-held video camera, where intentionality is very obvious), hence achieving the Reflectionist goal of apparent nonselectivity.
In particular, the apparatus provided a means of taking pictures of representatives of establishments that place customers under surveillance, in such a way that those representatives could not determine whether or not such pictures were being taken (just as we never know whether or not a department store surveillance camera is actually capturing an image of us at any given time).
WearCam comprised a computer system that was worn on the body, rather than carried, and a display means that left both hands free. A wireless connection to the Internet provided offsite backup of all image data, facilitating another aspect of the Reflectionist philosophy-namely, as far as destruction is concerned, to put the pictures beyond the reach of totalitarianist officials. Just as an individual cannot rob a bank and then destroy the video record (because the video is recorded or backed up offsite, or is otherwise beyond the bank robber's reach), my apparatus of détournement (see Fig. 2) put the images beyond the destructive reach of members of the establishment, because of the Internet connection, which allowed for offsite backup of all images at various sites around the world.
An advantage of transmitting images to remote locations is the possibility of having multiple processors work together at various remote sites to enhance the images by regarding each image as a collection of photometric measurements and combining these measurements together to reduce noise, extend dynamic range and tonal resolution, and increase spatial resolution and extent. In one such enhancement, I programmed the computers to use my algorithm to combine images together into a seamless photometric composite (Fig. 3), which provided a still image as a visual record of my gaze pattern.
(Note the irregularly shaped image boundary as well as the exceptionally high definition, often in excess of that attainable by photographic means.) My mathematical framework for this processing [28] has been successfully implemented on a large number of computers working in parallel, with a negligible amount of inter-processor communication.
More recently, the advent of the World Wide Web (WWW) facilitated my Wearable Wireless Webcam (1994) and the principle of offsite (off-body) backup was further enhanced. Once the image is distributed via the WWW, it is further beyond the destructive powers of department store security guards and the like, as I no longer know how many copies of my transmitted pictures might have been made. Evidence that might, for example, show that a department store has illegally chained shut its fire exits is not only beyond the store's ability to seize or destroy, but is also within easy reach of the fire marshall, who, following my directions via cellular phone from the department store, need only have a standard desktop computer with WWW browser in order to see first-hand what my call pertains to.
WearCam-on-the-WWW thus extends this "personal safety" infrastructure and further deters representatives of an otherwise totalitarian regime from being abusive: on one hand, I have collected the indestructible evidence of hostile totalitarian actions, and on the other, my friends and relatives are quite likely to be watching, in real time, at any given moment.
This process is a form of "personal documentary" or "personal video diary." Wearable Wireless Webcam challenges the "editing" tradition of cinematography by transmitting, in real time, life as it happens, from the perspective of the surveilled (Fig. 4). Furthermore, because I am merely capturing measurements of light (based on the photometric image composite [29], which represents the quantity of light arriving from any angle to a particular point in space), which are then yet to be "rendered" into a picture, I may choose to leave it up to a remote viewer operating a telematic virtual camera to make the choices of framing the picture (spatial extent), camera orientation, shutter speed, exposure, etc. (Fig. 4b).
In this way I may absolve myself of responsibility for taking pictures in establishments (such as department stores) where photography is prohibited, for I am merely a robot at the mercy of a remote operator who is the actual photographer (the one to make the judgment calls and actually push the virtual shutter release button). In this manner, an image results, but I have chosen not to know who the photographer is. Indeed, the purpose of these personal documentaries has been to challenge representatives of the video surveillance superhighway who at the same time prohibit photography and video.
These personal documentaries, such as one I call ShootingBack [30], typically had two audiences - the audience to which I performed and another, remote, audience. Members of the remote audience knew they were an audience because they were entering a traditional "gallery." (Even though it was virtual in the sense that it was on the Internet, it was still traditional in the sense that the interaction was analogous to a real-world gallery or museum.) The other audience comprised those who were physically present in front of the WearCam apparatus (e.g. representatives of the surveillance superhighway and customers/patrons of their establishments).
Members of the physically present audience, at first, do not realize that they are an audience. On one level, they might be regarded as "enemy" (they are being "shot at" in the sense of "shooting back"), while on another level, the performance is directed at them - to educate them, teaching being an act of love and human compassion.
ShootingBack was a meta-documentary (a documentary about making a documentary). Since I am a camera, in some sense, I do not need to carry a camera, but in ShootingBack, I did anyway. This second camera, an ordinary hand-held video camera, which I carried in a satchel, served as a prop with which to confront members of organizations that place us under surveillance. First, before pulling the camera out of my satchel, I would ask store representatives why they had cameras pointing at me, to which they would typically reply "Why are you so paranoid?" or "Only criminals are afraid of the cameras." All this, of course, was recorded by my WearComp/WearCam apparatus concealed in an ordinary pair of sunglasses. Then I would open up my satchel and pull out the hand-held video camera and point it at them in a very obvious manner. Suddenly they had to swallow their own words. In some sense, ShootingBack caught "the pot calling the kettle black."
The following are experiments that I have conducted and purposely taken to the extreme in order to (1) illustrate a point and (2) experience reactions and observations first hand. It is not likely that the average reader would go to these extremes but some more subtle variations of these experiments will still provide similar insights or reactions. In the tradition of conceptual art, they are presented in the form of "recipes," or lists of instructions. Some of them are simple enough to allow motivated readers to repeat these performances.
This particular piece (see also Fig. 6) is called Maybe Camera - Who's Paranoid?
Another variation of "Maybe Camera" involved making a large number of these shirts, but putting a real camera and transmitter into one of the shirts (I had someone with a repeater in a backpack provide an uplink to my car parked outside the shop, which in turn wirelessly uplinked to the Internet) and having a large group wear the shirts on the surveillance superhighway. Figure 7 depicts me with some family members wearing "Maybe Cameras."
"Probably Camera" and "Maybe Camera" can be worn together of course, since one uses the front of the body, while the other uses the back. "No Camera": This conceptual piece involves video time-delay [31], to symbolize the disjointedness between cause and effect that video recording creates:
In My Manager, I am metaphorically merely a puppet on a "string" (to be precise, a puppet on a wireless data connection). I might, for example, dutifully march into the establishment in question, go over to the stationery department, select a pencil for purchase, and march past the magazine rack without stopping to browse through the magazines, because I am not permitted by "my manager" to stop and browse. In this example, I have been sent on an errand to purchase a pencil for a higher and unquestionable authority. When challenged by the department store's clerks or security guards as to the purpose of the cameras I am wearing, I indicate that what I am wearing is a company uniform and that my manager requires me to wear the apparatus (the uniform) so that she can make sure that I do not stop and read magazines while I am performing errands on company time. Sometimes I remark: "I trust you, and I know you would never falsely accuse me of shoplifting, but my manager is really paranoid, and she thinks shopkeepers are out to get her employees by falsely accusing them of shoplifting" [33].
Just as representatives in an organization absolve themselves of responsibility for their surveillance systems by blaming surveillance on managers or others higher up their official hierarchy, I absolve myself of responsibility for taking pictures of these representatives without their permission because it is the remote manager(s) together with the thousands of viewers on the World Wide Web who are taking the pictures.
The subjects of the pictures - for example, department store managers, who had previously stated that "only criminals are afraid of video cameras" or that the use of video surveillance is beyond their control - either implicate themselves of their own accusations by showing fear in the face of a camera or acknowledge the undesirable state of affairs that can arise from cameras that function as an extension of a higher and unquestionable authority.
If their response is one of fear and paranoia, I hand them a form, entitled RFD (Request for Deletion) which they may use to make a request to have their pictures deleted from my manager's database (I inform them that the images have already been transmitted to my manager and cannot be deleted by me). The form asks them for name, social security number and why they would like to have their images deleted. The form also requests that they sign a section certifying that the reason is not one of concealing criminal activity, such as hiding the fact that their fire exits are illegally chained shut.
It is my hope that the department store attendant/representative sees himself/herself in the bureaucratic "mirror" that I have created by being a puppet on a (wireless) "string." My Manager forces attendants/maintainers/supporters of the video surveillance superhighway, with all of its rhetoric and bureaucracy, to realize or admit for a brief instant that they are puppets and to confront the reality of what their blind obedience leads to.
An early version of
Cyborgian Primitive involved my growing my hair through
fine mesh in
a skull cap and then "locking" it on the other side (hair
locking may
be accelerated by teasing in bee's wax to cause the hair to
tangle
together permanently). After I used conductive/metallic hair dyes
(to
help make my hair form part of a ground-plane for a transmitter),
my hair
was sufficiently "damaged" to lock quite easily. The skull
cap formed a
substrate upon which other devices could be mounted. In
this manner, I could
not reasonably be asked to remove the apparatus,
because that would require
shaving off my hair. This necessary
subversion of the body provided a
reasonable barrier to requests by
others that the apparatus be removed.
A more recent variant of Cyborgian Primitive depended on modifying the brain rather than the body. I based these experiments on something I have called "mediated reality" and proposed as a method of conducting scientific experiments in visual perception, as well as for prosthetic purposes [34]. As a prosthetic, the apparatus I describe in Fig. 2 of an MIT technical report [35] allowed me to computationally augment, diminish or otherwise alter the perception of reality for the purposes of attaining a heightened sense of awareness, seeing better or compensating for a visual deficiency that cannot be corrected with ordinary (pure-refractive optical) eyeglasses. Other researchers have experimented with the re-configuring of visual reality (Stratton experimented with optical upside-down glasses [36] and Kohler [37] and Dolezal [38] with left-right reversing prism glasses), but what is unique about my mediated-reality approach is that it is based on computational apparatus rather than optics (e.g. lenses, prisms and mirrors). Thus, my visual experience can be recorded and transmitted to remote locations, thus allowing others to augment, diminish or otherwise alter my perception of visual reality.
As have other scientists, I found that an adaptation to the apparatus occurred and that, after some time, I developed a dependence on the apparatus. Removal of the apparatus would result in my inability to see properly, as well as sensations of nausea, dizziness and disorientation. With this deliberate modification of the visual system, involving the development of alternate neural pathways through the process of certain kinds of very long-term visual adaptation, one may attain a permanent or semi-permanent bonding with the apparatus, in the sense that others cannot reasonably ask that it be removed. In the spirit of Reflectionism, WearCam was made to function as a true extension of the mind and body, as a third eye (or second pair of eyes, in the case of some two-camera systems I have described in my MIT technical report [39]).
Beyond the fact that a totalitarianist asking that the device be removed is asking the wearer to violate or subvert his or her own body, there is also the obvious legal responsibility the totalitarianist must accept for the prospect of the wearer's abrupt exposure to his or her original, or natural, neural pathways and the possibility of any brain damage or onset of flashbacks that might result from a sudden re-instantiation of the old (temporarily or semi-permanently weakened) neural paths.
Thus, when asked to remove the apparatus, if in fact it even could be removed (e.g. if it were not permanent or semi-permanent), one might merely present the totalitarianist attendant with a form to sign accepting all responsibility for any damage. This use of forms (e.g. an individual presenting officials with forms) is itself a Reflectionist gesture.
I recently used a joint mental and physical bonding (permanent/semipermanent head cap) in a self-ownership piece called Primitive Identity. In this piece, I defined myself as self + prosthetic device in all manner of official portraiture (e.g. Fig. 8), regardless of any requirements that eyeglasses and the like may not be worn during such portraiture.
As Foucault notes, it is not essential that the guard in the tower be watching a particular prisoner, or even that there be a guard in the tower; it is only necessary that the prisoner not know whether there is a guard watching in the tower. Similarly, to subvert Panopticon, it would not be essential that the guard be watched, but just that there be a possibility that the guard could be spotted by a "prisoner" at some time.
To this Diffusionist end, I have created a wireless communications infrastructure capable of supporting a networked community of hobbyists wearing a similar apparatus. During one performance piece, I, together with a group of others willing to participate, went out shopping one day, wearing such apparatus (thus, those at the department store needed to confront not just one, but many of us). The picture I took of this group was of such popularity that we recently re-enacted the event (Fig. 9).
Part of my Diffusionist goal is enhanced by finding everyday uses for wearable cameras---for example, cameras that automatically recognize faces, for individuals with visual or memory disability [40] (we all suffer from difficulty remembering faces), as well as wearable, tetherless computer-mediated reality for the public at large.
While one might be inclined to think that the inevitable commercialization of this invention may mark the détournement of this détournement, Diffusionism is put forth as a détournement of a détournement of a détournement (as in the equation Diffusionism = détournement3).
To this end, my goal is to turn WearCam into a useful and commercially viable everyday object that can help us see better, avoid getting lost (automatic directions combined with object recognition, global position systems [GPS] and video overlays), and remember names and faces better. Thus, these very utilitarian applications of WearCam may serve as a détournement of utilitarianism itself.
2. Michael Schneider, "In Baltimore, Big Brother Moves In," The Detroit News Home Page (20 Jan. 1996), http://www.detnews.com/menu/stories/32681.htm .
3. LynNell Hancock, Claudia Kalb and William Underhill, "You Don't Have To Smile," Newsweek (17 July 1995) p. 52.
4. Daniel Golden, "The Face of the Future," The Boston Globe Sunday Magazine, (30 June 1996). For related information see /wearcam/ biometrics_in_human_services.txt .
5. Simon Davies, "Privacy International Calls for CCTV Debate," ABC News 20/20 (8 September 1995), /wearcam/ privacy_forum_digest_on_CCTV.html .
7. Ali Azarbayejani, "Smart Spaces," http://www.siggraph.org/conferences/siggraph96/core/conference/bayou/b.html, 1996.
8. David Rokeby, "Camera-Based Performance Spaces" (1982).
9. In Rokeby's earlier system the user is in a dance/performance space---that is, a space where one expects to be observed, as opposed to a private space where one might have the desire not to be observed.
10. Andrew Wilson, Aaron Bobick, Lee Campbell, Elvis the Monster, Jim Davis, Freedom Baird, Stephen Intille, Arjan Schutte, Claudio Pinhanez, and Yuri Ivanov, "Kids Room" (1987), http:// www-white.media.mit.edu/vismod/demos/kidsroom/ .
11. Sr. Consumer Correspondent Hattie Kauffman, CBS Good Morning (television program), 3 July 1996.
13. MIT Media Laboratory, "Augmented Smart Chair," http://www-white.media.mit.edu/vismod/demos/smartchair/
14. Barry Winfield, "Buick Riviera," Preview (January 1994) pp. 124--127.
16. Joe Constance, "Nowhere To Hide: Holographic Imaging Radar May Soon Be Uncovering Hidden Dangers at U.S. Airports," http://www.ingersoll-rand.com/compair/octnov96/radar.htm.
18. This issue was discussed on an E-mail list moderated by Lauren Weinstein, "Privacy Forum Digest" (28 October 1996).
19. By self ownership, I mean that the same protections (e.g. copyright) governing the fruits of our labor (that which we intentionally put forth as a commodity) could also be applied to aspects of ourselves, such as our physical appearance, and other information that we generate unintentionally, just through our natural existence..
20. Michel Foucault, "Discipline and Punish," Alan Sheridan, trans. (New York: Vintage, 1977). Originally published as Surveiller et punir (Paris: Gallimard, 1975).
21. Elisabeth Sussman, "On the Passage of a Few People through a Rather Brief Moment in Time," in The Situationist International 1957-1972, exh. cat. (Cambridge, MA: MIT Press, 1991) p. 127. This exhibition catalog includes essays, illustrations and artistic documents for a retrospective held at the Pompidou Center, the Inst. of Contemporary Arts in London and the Institute of Contemporary Art in Boston from 1989 to 1990.
22. Tom Ward, "The Situationists Reconsidered," in Douglas Kahn and Diane Neumaier, eds., Cultures in Contention (Seattle, WA: The Real Comet Press, 1985.)
23. "[Détournement] . . . is the art of appropriating common objects or images from their usual cultural contexts and resituating them in an incongruous, disturbing, and disorienting fashion in order to confront, question, or challenge society's stereotypes or biases." From W. Ted Rogers, in Sunil Gupta, ed., Disrupted Borders: An Intervention in Definitions of Boundaries (London: Rivers Oram Press, 1993). Détournement is short for "détournement of preexisting aesthetic elements."
24. Steve Mann, "Wearable Computing: A First Step toward Personal Imaging, IEEE Computer, 30, No. 2 (Feb. 1997) pp. 25--32. Also published as the feature article of the February 1997 entry in http://computer.org/computer/backissu.htm.
25. Steve Mann, "Existential Technology," unpublished manuscript.
26. I first developed WearComp in the 1970s as a "photographer's assistant" for controlling sources of illumination. This effort evolved into a new system of creating expressive images based on the linearity and superposition properties of light.
27. S. Mann, "Wearable Wireless Webcam," 1994; currently at /wearcam.
28. S. Mann and R.W. Picard, "Video Orbits of the Projective Group: A Simple Approach to Featureless Estimation of Parameters," IEEE Transactions on Image Processing, 6, No. 9 (September 1997). Also publishing as Tech. Report 338, MIT Media Lab, Perceptual Computing Section (Cambridge, MA: 1995).
29. Steve Mann, "`Pencigraphy' with AGC: Joint Parameter Estimation in Both Domain and Range of Functions in Same Orbit of the Projective-Wyckoff Group," IEEE International Conference on Image Processing (ICIP 96) (Lausanne, Switzerland: September 1996). Also published as Technical Report 384, MIT Media Lab, (Cambridge, MA: December 1994).
30. Steve Mann, "ShootingBack: Personal Imaging in Personal Documentary," 1996, unpublished manuscript. See also/wearcam/shootingback.html. .
31. Other artists have also experimented with video time-delay but in different contexts. For example, Dan Graham uses video time-delay together with mirrors, etc., to create a delay between cause and effect. His video feedback involves both senses of the word "feedback": (1) the cameras "sees" the screen, which is displaying the output from the camera, and (2) the users who see themselves on the screen adjust their behavior according to this psychological "feedback."
32. Both Stelarc and Elsenaar explore body control systems that use electrical stimulation to cause their muscles to move in response to an external input. See Stelarc's official web site, Australia, 1997,http://www.stelarc.va.com.au/ http://www.merlin.com.au/stelarc/ ; and Arthur Elsenaar, 1997, http://www.desk.nl/~acsi/WS/artists/elsenaar.htm and /wearcam/previous_experiences/arthur_elsenaar/
33. There are well-documented cases where shopkeepers have falsely accused their customers of shoplifting, so my assertion is not as absurd as it might seem. It is quite reasonable that individuals keep their own video records of their experiences in shops, as a sort of "black box" recorder in case such an accusation should arise. In some cases, officials have raped or murdered patrons of their establishments, so it seems reasonable that officials should not be the only ones to have video records (e.g. that they could erase). In one well-known murder case: "On March 16th, 1991, 15 year old Latasha Harlins entered a Korean owned grocery store to purchase a carton of orange juice. Soon Ja Du, the store owner, accused her of shoplifting even as Latasha attempted to pay for the juice. After a struggle in which Du tried to grab her book bag Latasha placed the juice back on the counter. As Latasha turned to go, Du shot her in the back of the head, killing her." From documentary video "Hands on the Verdict: The 1992 L.A. Uprising," produced for Deep Dish T.V. Coordinating producers: Liz Canner and Juloiea Meltzer. See also "Korean Grocer Convicted in Shooting," New York Times (12 October 1991).
34. S. Mann, "Mediated Reality," Tech. Report 260, MIT Media Lab, Perceptual Computing Section (Cambridge, MA: 1994).
36. George M. Stratton, "Some Preliminary Experiments on Vision," Psychological Review 1896.
37. Ivo Kohler, "The Formation and Transformation of the Perceptual World," Vol. 3 of Psychological Issues, (Vienna: International Univ. Press, 1964) Monograph 12.
38. Hubert Dolezal, Living in a World Transformed, Cognition and Perception Series (Chicago, IL: Academic Press, 1982).
40. Steve Mann, "Wearable, Tetherless Computer-Mediated Reality: WearCam as a Wearable Face-Recognizer, and Other Applications for the Disabled," Tech. Report 361, MIT Media Lab, Perceptual Computing Section (Cambridge, MA: 2 February 1996). Also published in AAAI Fall Symposium on Developing Assistive Technology for People with Disabilities, 9-11 November 1996, MIT
41. D. Hockney, Hockney on Photography: Conversations with Paul Joyce (London: London Cape, 1988).
42. S. Mann, "Compositing Multiple Pictures of the Same Scene," in Proceedings of the 46th Annual IS&T Conference (Cambridge, MA: Society of Imaging Science and Technology, 1993).
A PostScript version of this article is also available.