(Presented at UKAIS Conference, Easter 1997)
Abstract
What are the consequences of capturing information about people in databases
and information systems? We argue that, although more and more data
is being collected about people's actions and transactions, the result
is not that an increasingly full picture of individuals' lives is emerging
in the databases, but that, since such collection is not coherently organized,
multiple partial images are accumulating which misrepresent individual lives
and have a de-centring effect on them. People cannot be properly known
through these images, nor could they be even if the images were unified in
some way, since the images are controlled by the collectors, not the people
themselves. We outline the stages involved in collecting and using
personal data, point out the inevitable misrepresentation involved in such
practices, argue that it is important to increase awareness of them, and
suggest possible responses, ranging from acceptance to more oppositional
strategies.
"Hey there, Mr. Average,
You don't exist, you never did,
Hiding in shadows –-
Persons unknown"1
1. PERSONAL
DATA: A HALL OF MIRRORS
TOP
MAKING DATA REPRESENTATIONS OF PEOPLE
OPPOSING PERSONAL CODING
REFERENCES
Mark Poster, in The Mode of Information, argues thus about the dangers inherent
in database representations of people:
‘"I contend that the database imposes a new language
on top of those already existing and that it is an impoverished, limited
language, one that uses the norm to constitute individuals and define deviants."
(p95)2
He goes on to state that further analysis shows the population participating
in its own self–constitution as subjects under diffuse, normalizing
processes built into database systems. His chief worry about databases
is not so much that they invade privacy as that they create replacement selves
which may obscure and harm the originals:
‘"We see databases not as an invasion of privacy,
as a threat to a centred individual, but as the multiplication of that
individual,
the constitution of an additional self, one that may be acted upon to the
detriment of the "real" self without that "real" self ever being aware of
what is happening. .... The innocuous spread of credit card transactions,
today into supermarkets, tomorrow perhaps into classrooms and homes feeds
the databases at ever increasing rates, stuffing ubiquitous computers with
a language of surveillance and control. Rather than the motto "all
information at all places at all times" an oppositional strategy might better
follow Lyotard in his conclusion to The Postmodern Condition: "give the public
free access to the memory and data banks".’" (pp 97–- 98)
2
We would like here to take this a little further Poster's concern
about the ‘"multiplication" of individuals in databases
and sketch out what an "oppositional strategy" might be. We start
by arguing that computer representations of people do in fact constitute
a threat to the centred individual, precisely because they multiply individuals
- not merely by producing ‘"an additional self" as Poster
suggests, but by producing multiple partial and uncoordinated selves.
It is not that an alter ego is produced in a computer representation, but
that many distorted and fragmentary reflections are produced in different
information systems, trapping individuals in a ‘"hall of mirrors".
We do not suggest that individuals are naturally or ideally ‘"centred",
in the sense of being coherent emotional and cognitive wholes, recognizing
the force of the arguments against such a conception. Sampson, for
instance, reviews a number of challenges to the notion of the individual
person as"psychology's subject", in particular the attacks
on the primacy of the subject coming from critical theory and deconstructionism.
Critical theory invites us to consider the possibility that the ’"individual
person" is a character designed primarily to serve ideological (bourgeois)
purposes; while deconstruction denies any primacy to the subject or author,
arguing instead that the sense of self–-presence is a delusion. The
deconstructionist attack on the notion of the individual self as a centre
of awareness concludes that persons are not centres of awareness, but are
decentred by their relations to a shifting and indeterminate symbolic order.
3
But as the symbolic order becomes increasingly fragmentary and kaleidoscopic,
as images and representations multiply –- in databases and in many other
media –- this process of decentring intensifies, threatening the integrity
of the individual. The question is whether it can be or ought to be
slowed down or reversed.
Giddens has challenged as ideological the view that ‘"the end of the
individual" is either a desirable or inevitable movement of contemporary
social change, and insists that ‘"the de-centring of the subject
must not be made equivalent to its disappearance’"
4. He argues against the conceptual elimination of the subject,
and in favour instead of promoting a recovery of the subject
5.
What can be done in the field of information systems to promote such a recovery?
In so far as information systems, especially as practised, draws from positivist
and structuralist traditions, it tends to overlook or dissolve the subject.
There are general questions to be asked about the roles or actions of subjects
in and of information systems, and more particular questions here about subjects'
relations to representations of themselves in information systems.
MAKING DATA REPRESENTATIONS
OF PEOPLE
TOP
PERSONAL DATA: A HALL OF MIRRORS
OPPOSING PERSONAL CODING
REFERENCES
The first point to make is that creating and using data representations of
persons changes the way we know them - and for the worse. We come to
know about people in terms of their observed or interpreted behaviours and
the records of some transactions they have been involved in, rather than
knowing them as persons in their own right.
We know about them in many different ways at the same time. There is
a continuous rise in the number of ways in which our actions and transactions
are recorded, and in the number and variety of individuals' (trans)actions
from which data is captured. Examples of these processes include the
census, use of credit cards, the rise of closed circuit TV
6, BT data warehousing7
and supermarket "loyalty" schemes8.
The recording of our actions and transactions is always done for some purpose,
and each of these purposes entails a view of our behaviour and necessarily
a reduction of it. General purposes for collecting data about people
include :—-
- to perform an immediate process (e.g. keeping track of the number
of persons in a venue for safety reasons)
- for some later use (e.g. targeting specific customers to entice
them to buy new products)
- For general monitoring - just in case a need arises (e.g. security
services)
In order to record an aspect of people's actions, a number of steps
have to be taken:—-
- categorization (of actions and people)
- capture (of information)
- coding (into the categories)
- database entry
The important questions individuals should be asking, for each of these steps,
are: Who carries them out, and for what purposes? As far as categorization
is concerned, we agree with Rowe when she writes, "Power is always
about who does the defining and who accepts the definitions."
9 A crucial point here is that the categorizations employed
are imposed upon those persons who are the generators of the data that is
captured — there is a power imbalance designed into the (trans)actions.
This power imbalance is magnified because the persons to whom captured data
refer often do not even know that they are accepting these categorizations,
since they are not made clear to them. Furthermore, since the purposes
of different agencies collecting information about us will generally be different,
no coherent view of us is being built, even if all the information were to
be collected together.
As data capture becomes increasingly automated (e.g. CCTV and credit card
data) or is carried out by "operatives" using "invisible" surveillance technologies
(as Lyons10 has put it), the
purposes for which the data is being gathered become increasingly obscure,
or turn into an apparently arbitrary exercise of power (e.g. taking photographs
and car registration numbers at demonstrations). When the data collection
becomes detached from immediate purposes, problems of interpretation increase.
The use of photographs or video as objective base data, for instance, is
fraught with difficulties of identification and interpretation
11 and does not in general provide the unquestionable depiction
of events that some of the proponents of these approaches might want us to
believe.
Once captured, how is the data used? This is the greatest area of concern,
because it is here, away from the point of collection, that the data is transformed
to serve the purposes of its collection, and where there is most danger of
misuse or abuse of information. Uses of personal data typically involve
these steps:—-
- matching and fitting to categories (and approximations, which are
inevitable given the nature of the technologies employed
12)
- aggregation, summary, filtering and normalization
- computer-mediated dialogue (in processes of querying, extracting
and presenting).
The same pressing questions apply here as before: Who carries out these steps,
and for what purposes? Given that the persons to whom the captured
data refer are almost never privy to the subsequent use of that data, there
is another designed power imbalance here, in the use of personal data.
There is a further distancing between source and use of data arising out
of the fact that, frequently, data is captured by "operatives" (human &/or
electronic), but used by "managers"; the managers are likely to have been
involved in defining the categorizations used in data capture, and the "operatives"
not at all. Categorization and use are thus brought together, but out
of sight of the producers or collectors of the personal data.
The inevitable result of the processes of categorization, capture and use
of data as outlined above is a fragmentary picture of persons —- or
rather several such, since the fragments are not part of a single picture.
If multiple agencies capture data about people for many different purposes,
a single person will give rise to many different representations on many
databases, with each representation capturing separate aspects of that person's
life. The representation of a person held by the Inland Revenue, for
instance, will have little overlap with that held by British Telecom, apart
from name and address details. No simple "data image" of a person can
be produced this way, since there is no single representation in existence,
nor in prospect (nor arguably even desired by any of the collecting agencies,
who are less interested in persons as such than in particular controllable
behaviours). What we have about persons in the various databases is less
an image than pieces of various mosaics, or perhaps only dust.
The consequence of these multiple codings is that a person remains always
unknown in extenso, and only ever known in relation to the purposes of the
coders, for whom the person is replaced by a code replica. What the
persons represented think about being so replaced is hard to establish: perhaps
they do indeed collude with it, as Poster suggests; perhaps the processes
are so piecemeal and obscure as not to impinge on them. Considering
the situation in our own immediate case, our response combines rejection
of coding of persons as a meaningful process with incredulity that we put
up with it.
OPPOSING PERSONAL CODING
TOP
PERSONAL DATA: A HALL OF MIRRORS
MAKING DATA REPRESENTATIONS
OF PEOPLE
REFERENCES
Pointing out the piecemeal and covert processes at work in personal data
capture will not stop it happening, but may illuminate the need for a response
and clarify possible forms such a response might take. A response becomes
necessary when we realize personally, for ourselves and for others, that
collection of data about people is not simply a mundane operation, essentially
harmless, nor even a concerted attack on individual privacy, but is, more
insidiously, an uncoordinated process which progressively fragments personal
coherence.
How might this fragmentation be countered? If persons are reduced in
databases to code replicas of their habits, characteristics, and circumstances
as judged relevant to the collectors and users, might we be able to give
individuals the chance to say more (or less) about themselves? Opportunities
could be offered for people to examine their representations, to correct
errors, to give a fuller account of themselves, or perhaps to represent themselves
not as individuals at all, but in the context of their belonging to particular
communities.
In general, several responses are possible to the coding of people:—
- Acceptance - this is what all of us do to some extent, once we
know about data capture of our (trans)actions. We use a telephone at
home, or a credit card, or walk into stores with CCTV cameras. This
response is generally justified by the view that resistance is pointless,
or that the benefits that accrue to us outweigh the disadvantages.
(This position becomes less easy to adhere to if you accept the foregoing
analysis.)
- Raising awareness - if people are not aware of the processes of
data capture and use, then they are deprived of the ability to choose any
of the responses. Those who know more of these processes could share
their knowledge. People engaged in the design of systems of data capture
and coding, such as computing professionals, we would argue, have a moral
obligation to alert the subjects of these systems that their actions are
now being traced.
- Negotiation –- one might inquire about the purposes behind
the data capture and subsequent coding of persons, and seek to negotiate
the representations actually stored. This presupposes the possibility
of dialogue between data collectors and users and data generators.
It is a co-operative stance but contains the expectation of a reply.
If a reply is not forthcoming, other responses may follow.
- Refusal - this can take the form of non-cooperation with the process
of data capture, e.g. not providing non-essential data to data capture "operatives",
or by restraining the usage with legal contracts
13. With technological data capture devices (e.g. cameras) it
can take the form of obscuring one's face or other identifying characteristics.
Another form of refusal involves laying false trails, e.g. giving incorrect
details on vouchers. Refusal is the clearest "oppositional" strategy,
likely to appeal to (among others) those most concerned by the disintegrative
effects of the unchecked accumulation of personal data.
Our analysis leads us to conclude that any "full" or "true" representation
of persons in databases and information systems would be very hard to achieve.
So long as categorization and collection are controlled by others, it will
be impossible. If people control their own representations, coherent
images would be more possible, though even a self–-coding will be inevitably
incomplete. In any case, comprehensive representations of people in
databases are undesirable if only because they could fall under some uncongenial
form of external control.
The best general solution to the problem of personal data collection might
take the form of specific negotiated representations with local encryption,
with the key to the encrypted data left under the control of the person whom
the data represented. This is not as infeasible as might be thought
- the European Union is already investigating such approaches
14. Such an arrangement would have the advantage that, while
people would have a say in how much and how far they are known in the data,
unaccountable accumulations and transfers of fragmentary personal data would
be kept in check. To preserve the integrity of persons in the community,
it is important that they remain relatively unknown in the databases.
Throwing the databases open is a good strategy, but keeping local control
over their definition and usage is better still.
REFERENCES
TOP
PERSONAL DATA: A HALL OF MIRRORS
MAKING DATA REPRESENTATIONS
OF PEOPLE
OPPOSING PERSONAL CODING
- Poisongirls (Subversa/Poisongirls) (1995) Persons Unknown.
Cooking Vinyl/BMG Music Publishing.
- Poster M. (1990), The Mode of Information. Polity
Press. Cambridge.
- Edward E Sampson (1989). The Deconstruction of the Self in
Texts of Identity ed J Shotter & K J Gergen, Sage. London.
p 14
- Anthony Giddens (1979). Critical Problems in Social Theory,
Macmillan. London. p45
- ibid, p44
- John Naughton (13th November 1994). Smile, You're on TV,
Observer
- Emma Mansell-Lewis (7th December 1995). Sense and Sensibility,
Computer Weekly
- Michael Durham (7th April 1996). Your Supermarket Needs You,
Observer
- Dorothy Rowe (1987). Beyond Fear, Fontana/Collins
. London, p42
- David Lyons (1994). The Electronic Eye: The Rise of Surveillance
Society, Polity Press. Cambridge.
- Chris Mihill (21st December 1995). Photo Test Casts Doubt
on ID Plan. Guardian.
- William Kent (1978). Data and Reality, North Holland.
Amsterdam. passim
- Dan Glaister (27th March 1996). Mail Check, Guardian.
- Dan Glaister (14th February 1996). Little Brother,
Guardian.