Presentation for my talk at the 2011 ITiCSE (Innovation and Technology in Computer Science Education) conference in Germany.
My paper and talk critique the most common way of teaching the social and professional issues part of the computer science curriculum, which is to focus on the ethical evaluation of "impacts" caused by computing technology.
This presentation argues that this approach has a number of drawbacks. First, it is based on a technological deterministic style of social explanation that has been in disrepute in the academic social sciences for decades. Second, it uses an algorithmic approach to ethics that simplifies the social complexity and the uncertainty that is the reality of socio-technological change.
16. It is not hard
to see
the attraction of this
approach
for
f us
computer professors
17. The “many gray areas
of computer ethics are
often frightening …
to professors
who are worried
about how to answer
things of which
they themselves are
unsure.”
19. Some have indeed argued that
using this algorithmic methodology
provides a sense of security and confidence
for nervous CS faculty teaching an SPI course
20. I am going to argue that both
this algorithmic methodology
and
the theoretical understanding of the relationship between
technology and society that it is based upon
are d
deeply fl
l flawed.
d
21. I will try to convince you that
we need to move away from
y
the ethical evaluation of impacts
and instead
emphasize
h i
the social context of computing
22. In particular, the way we teach SPI material
needs to integrate the decades-old insights
of researchers in the philosophy, history, and sociology of technology
that emphasizes a very different approach to SPI.
23.
24. The common way of seeing technology is that
it is akin to a cue ball
k b ll
impacting or altering the rest of society
25. In this perspective
key technological inventions
have transformed the world.
h f d h ld
Thus new technologies
need to be analyzed to understand the
wide changes they will enact.
26. This approach to technology
is generally referred to as
Technological Determinism
T h l i lD t i i
29. It is understandable
why computer professionals find
technological d
h l i l determinism attractive.
i i i
We are the people helping to invent
new technologies
30. It feeds our to be
clear desire socially
relevant
31. … and our desire
to believe that we
computer geeks k
are the driver of
social change, and
g ,
not politicians,
business people, or
celebrities.
32.
33. Most current historians and sociologists of technology
firmly reject technological determinism
because it is
theoretically inconsistent
and
empirically under-supported
p y pp
34. The well-established academic field of
science,
science technology and society (STS) studies
(that began in the 1960s)
has time and time again found that when
examined carefully
i d f ll
most technologies rarely have had the effect
that was expected
or
had the transformative impact people claim.
35.
36.
37. economy
history
technology
culture
society
politics
38.
39. 1. Empirically f l
1 E ll false
2.
2 Not used by STS research community
3. Naively focused on functional capabilities
y p
40. functional capabilities?
Most technological deterministic impact prognosticators do their work
by looking at the functional capabilities of a given technology
and then imagining the impact of those functions.
41.
42.
43.
44.
45.
46. In all these cases
– and practically any other set of prognostications and
impact evaluations than begin from an unquestioned belief
that the functional capabilities of a technology (i.e., the
means) do what is promised (i.e., achieve their ends) –
the expected social impacts
ended up being wildly wrong
because the prognosticators
believed in a
naïve technological determinism.
47.
48. The introduction of household technology
did not end up creating,
in the words of Ruth Schwartz Cowan,
less work for mother,
but
in fact
more work
because of a series of social changes that could not h
b f i f i l h th t ld t have b been
predicted if one limited one’s analysis just to the functional
capabilities of the household technologies.
49. Efficient internet search-engines
have not
resulted in people with
too much knowledge;
instead,
i d
unpredicted changes in how people interact
with words and even possibly cognitive decline
due to the brain s plasticity
brain’s
have arguably resulted in the exact
opposite consequence
50. The introduction of anti-lock disc brakes
have not
reduced accidents at all,
because drivers tend to drive faster and tailgate more
closely due to the improved braking technology and
also partly because of increases in the intensity of
traffic due to unexpected changes in urban
geography.
51. Depending upon how failure is defined,
anywhere from 50 to 90% of ERPs have
been categorized as failures.
Others
O h report over 70% achieve few of
hi f f
the objectives expected of them.
There are even examples where bankruptcy trustees
Th l h b k
have blamed the adoption of ERP software for the
failure of the business as a whole.
52.
53. The first step
Then we should take in our
Social and Professional Issues courses
is to communicate how rarely
technologies achieve their promise
promise,
and indeed,
how many do the opposite.
54. The revenge effect of technology
is extremely well documented
yet
it is uncovered in any of the
computer ethics textbooks examined f
hi b k i d for
this presentation.
55. One way to achieve this goal
would be by beginning the SPI course
with examples and readings
on how certain vital technologies had little
impact on some societies,
or
on how certain technologies were strongly
modified and differently adapted in
y p
different cultures and countries.
57. In this approach,
one looks at how technologies are
researched,
invented,
financed,
financed
developed,
adopted,
marketed,
and propagated
within a very complex system
generally referred to as society.
58. In other words,
our SPI courses should
look more like a
historical sociology course
and a lot less
like a philosophic ethics course.
60. The reason revenge effects occur
is due to the fact that
“socio-technological transformation
socio-technological
is a highly complex process
which involves many uncertainties ”
uncertainties.
61.
62. But in moral philosophy
p p y
and computer ethics in particular
uncertainty is underappreciated.
This is an important issue
since
substantive moral theories
such as deontology and utilitarianism
h d l d ili i i
require clear information
about
effects in order to
make ethical judgments.
63. Typical problems or dilemmas
yp p
for which macro-ethical approaches are
applied are most often done in a
context of complete knowledge.
This is appealing for computer scientists,
who often work with problems
modeled by idealized abstractions
for which complete knowledge is possible.
Function PerformEthicalEvaluation()
If (you do action x) then
y people will be harmed
but z people will be benefited
End If
Return EthicalEvaluation(y,z)
End Function
64. The ethics of technology,
by contrast,
should be recognized
as residing
in
i a context
of at least partial
p
uncertainty or ambiguity.
66. the degree of uncertainty
is greater
for
f emerging t h l i
i technologies
and the more complex the technology
the more uncertain we are
as to the developmental
trajectory of a technology
67. o1 o2 o3 = different development trajectories
t1 = time one y = specific technology x = evaluating agent
teacher
student
journalist
68. The evaluating agent
must have knowledge
of the development trajectory
in order to morally evaluate it.
70. And if we guess one based on
functional capabilities we will
usually pick the wrong trajectory.
71.
72. It might be more helpful and
important to understand which
(this requires skills in sociology and history)
trajectories are more likely than
to apply a prescriptive ethical
j g
judgment to a single trajectory.
g j y
73.
74. In the early years of the 1990s, the
computers and society literature
provided ethical evaluations of a
variety of trajectories that never
materialized, such as
The disappearance of war
f
The workless society
The elimination of large corporations
Citizen control over all aspect of political life
76. This paper has argued that
moral evaluation in the
uncertain realm of
socio-technological
socio technological change
should only be tentative at best.
78. Rather than applying
big ethical theories
to l
t clear and well-known impacts,
d w ll k w i t
the focus should be on disclosing the
assumptions,
ti
values,
and interests
d
built into the
design,
implementation,
and use of technology.
79. We can in fact satisfy the CS-2008 SPI area
by guiding our students in the unpacking of the
normative assumptions of computer practice,
which in turn requires clarifying the complex web
of interactions that construct the different
trajectories a technology may take.
80. This
Thi approach is not about evaluating the
hi t b t l ti th
rightness or wrongness
of a technological practice
81. Rather, it is about opening up
the bl k box of
h black b f
technological practice
g p
for understanding …
82.
83. ACM recognizes that it is equally important for students to appreciate
the historical and social context as it is to perform ethical evaluation.
84. It is time for us to transform the
way we teach the SPI area, so it
is more in tune with the actual
ACM recommendations.
(as well as more in tune with approaches in
other disciplines studying technology)
85. Doing so would make
our SPI courses much more
focused on the
f d th
social contexts of computing
andd
significantly less focused
on i ethical evaluation.
its hi l l i
86. Randy Connolly
Dept. Computer Science & Information Systems
Mount Royal University, Calgary, Canada
y y, g y,
rconnolly@mtroyal.ca
Images from iStockPhoto and stock.xchng