The Responsibility Problem in Engineering

1
The Responsibility Problem in Engineering
Thomas M. Powers, Department of Philosophy Santa
Clara University Tpowers_at_scu.edu
2
2 main questions
P1. What does it to mean to be a responsible
engineer? P2. Does the notion of responsibility
vary over fields of engineering?
3
H.L.A. Harts Conception of Responsibility
P1. Causal Responsibility praise and blame P2.
Role Responsibility professional ethics P3.
Legal Liability P4. Moral Capacity as a
necessary condition
4
Causal Responsibility
PTo be praised or blamed is to be held causally
responsible for a good or bad outcome as the
agent or originator of the cause.
5
Causes Necessary or Sufficient Conditions?
P1. Logic of a necessary condition PIf the
effect, then the condition. P(Effect Condition)
P2. Logic of a sufficient conditon PIf the
condition, then the effect P(Condition Effect)
6
Role Responsibility
POne is held to be responsible in this sense if
one occupies a role in which the welfare of
people, the performance of a product, or the
functioning of a system is an explicit part of
the role (job, position).
7
Legal Responsibility
A function of causality, roles, and laws
PThe traffic accident as a case- study in the
logic of causes and responsibility
8
Moral/Rational Capacity
A necessary condition for other types of
responsibility
PWhen persons are held to be incapable of the
moral decision process, they cannot be held
responsible. PExamples Criminal insanity,
infancy, drugged dementia, etc.
9
Engineering Responsibility in General
Some Central Issues
PDesign PQuality Control PSystems
Theory POptimization and Risk PMarket Pressures
10
A Special Problem for Software Engineering
When is software sufficiently tested for release?
P It is well known that programs of even moderate
complexity are not verifiable within any time
frame of interest to humans. That is, many
programs cannot be proven "correct" or free from
bugs within the number of years in which they
would be in use. This limitation is of both
theoretical and practical importance it is
wrongheaded to demand of software engineers
programs that would be verifiable, and it is
futile to embark on testing projects that will
not come to fruition. B
11
Does the Atesting_at_problem show that software
engineers cant be responsible?
P Since most programs cannot be proven correct,
it may seem that this theoretical limitation
gives us a moral pass expending no effort at
software quality control seems to be the same as
expending maximal effort. In light of this
problem, we often turn to software testing as the
next-best solution. But how much testing is
enough? And what kinds of tests ought to be
done, so that we can honestly say that the
software professional (programmer, tester, or
team) is acting responsibly in releasing the
program to the customer?
12
3 Theories of Responsibility
PAbsolutist PUtilitarian (CBA) PKantian
13
For Further Research
PThe "Killer Robot" case http//onlineethics.org/
cases/robot/robot.html P"Professional
Responsibility The Role of Engineering in
Society" by S.P. Nichols and W.F. Weldon, Center
for Electromechanics, The University of Texas at
Austin http//www.me.utexas.edu/srdesign/paper/
PEngineering Ethics Website at TexasAM
http//lowery.tamu.edu/ethics/