The Best of Creative Computing Volume 1 (published 1976)
Computer Abuse: The Need for a Rational Perspective
ailments and using a standardized set of diagnostic criteria,
are less accurate in their diagnoses than computers
analyzing the identical criteria.
The explanation for this disparity in performances seems
to relate to several factors, including the doctor's state of
mind, and their reactions to extraneous aspects of the
patients' physical appearance and personality. The significant
point here is that, for routine, repetitive situations of
all sorts, where criteria/characteristic relationships are well
understood and empirically demonstrable, computers do
have legitimate, cost-effective application potential, and in
fact, may well out-perform their human counterparts.
(Computer auto diagnostics are a superior example of this
sort of application.)
While we're on the subject of human shortcomings, let
me bring up a couple of additional study findings of the
past several years which serve, I believe, to put the
computer into better perspective. For example, clinical
experiments strongly suggest that human analytical
processes tend to remain unchanged, even in the face of
extremely high rates of dysfunction. ln one test,(Goldberg,
American Psychologist, 1970), diagnosticians were given a
hypothesis about the interpretation of certain test results,
along with several corroborating case histories. Using these
criteria, the clinicians were given a series of test cases to
analyze, receiving immediate feedback regarding the
correctness of their evaluations. It took an 80% failure rate
to make them consider abandoning the criteria they had
been given and to search for a more suitable one.
Data such as this suggests that the human mind tends
not to be adaptive. Yet a computer can be programmed to
put reviewers on notice whenever a failure rate on a routine
procedure reaches an unacceptable rate, which we would
judgmentally set considerably lower than 80%. There is
clinical experience which suggests further, rather severe
limitations to man's cognitive and perceptual abilities.
Numerous studies tend to corroborate George Miller's
original hypothesis that the human mind can handle no
more than 7 2 different data elements at the same time.
Stafford and de Neufville state not only that it is impossible
for humans to make reliably consistent comparisons
between objects with more than one dimension, but that
human judgment, particularly in non-linear decision
situations, is clouded by personal psychological utilities.
(Systems Analysis for Engineers and Managers, McGraw
Hill, l97l.)
Thus, man is not without his shortcomings as an
analytical tool, as the supercilious HAL observed in Arthur
Clarke's "2001". In fact, one might go so far as to suggest
that man is a 'double-edged sword' whose potential for
social and political evil must be carefully proscribed. An
absurd analogy, and yet, of course, we do proscribe man's
potential for doing wrong through our laws. And so too, I
suggest, should we deal with the negative potentials of
Computers.
Basically, of course, when we talk about the regulation
of computers, we are really talking about controlling the
use and abuse of data. A variety of approaches have been
suggested for the regulation of data banks. Some people
have proposed that certain classes of personal data simply
not be automated (or even collected at all, for that matter).
Others have proposed that the collection, processing and
distribution of information be licensed by the government.
The most far-reaching, concrete action of this sort has been
the Swedish Data Act of 1974, under which no
computerized personal data file may be established or
maintained without a permit issued by a Data Inspection
Board, which also issues and enforces standards for file size
and content, security, disclosure, and data sharing.
It is clear that there are a variety of proposals for data
bank regulation which would be workable and effective.
Such proposals, however, fail to address a critical reality
which will become an increasingly important central feature
of the information age. This reality is that information has
economic value. There is ample evidence of this in the every
day life of organized social, political, and economic
enterprise. Institutions spend millions of dollars on reports
and studies to aid them in making correct decisions. The
sale of abstracts, digests, bulletins, news and technical
journals run into the billions of dollars. Mailing lists,
particularly when associated with socio-economic data, are
of enormous value for marketing purposes. Patent holders
sell or lease the rights to use their ideas. The "all news"
radio format is now the biggest money maker in the audio
broadcast industry. Institutions pay millions of dollars each
year to send employees to schools, seminars, and
conferences ... presumably for the knowledge which the
employees will acquire for use within the institution.
Of course, information has always had value, and has
always been bought and sold; but never on so large a scale
or in so large an amount as it is today. The primary reason
for this is the complexity of the modern world.
Decision-makers seek to minimize risk; with the increased
variety in our society, more and more data must be
gathered and analyzed in order to determine the
probabilities associated with a given decision. lt is clear that
this situation will continue to accelerate as individuals and
institutions seek to improve upon the sub-optimal decisions
of the recent past. Thus, the compilation, analysis, and
distribution of information will become a vital global
industry in the coming decade. Some have speculated that
it will become the dominant industry. Given these
circumstances, it would seem extremely desirable and
patently reasonable that we should immediately begin to
treat information as an economic commodity.
First, it should be pointed out that the courts have long
linked the protection of personal privacy to the legal
conception of property rights. In this context, the
unauthorized acquisition and/or dissemination of private
information has been treated under the laws pertaining to
trespass; while the unauthorized distribution of legitimately
acquired data is regarded as a breach of contract. (The new
Swedish Data Act established "Data Trespass" as a new
form of crime.) Thus, for most personal data, the law is
already well established with regard to the treatment of
information as a commodity.
[image]
