You are on page 1of 3

IT Ethics

Moral
Responsibility
for Computing
© Winterberg | Dreamstime.com

Artifacts:
“The Rules”
Keith W. Miller, University of Illinois at Springfield

I
n March 2010, the Poynter on Responsible Computing and of moral responsibility for these
Center for the Study of Ethics suggest changes for The Rules by artifacts, and to encourage individ­
and American Institutions emailing the coordinator. As the uals and institutions to carefully
held a workshop sponsored current coordinator, it’s my job examine their own responsibilities
by the US National Science Foun­ to circulate new suggestions and with respect to computing arti­
dation. An interdisciplinary group incorporate the accepted changes. facts. We do not claim that these
of philosophers, computer scien­ Here, I present a condensed ver­ rules are exhaustive; professionals,
tists, practitioners, and lawyers sion of the latest document, “Moral individuals, and organizations may
gathered to discuss “ethical guid­ Responsibility for Computing choose to take on more responsi­
ance for the research and applica­ Artifacts: Five Rules, Version 27.” bility than we describe here.
tion of pervasive and autonomous The reasons that the document
information technology” (http:// has gone through 27 versions A Working Definition of “Com­
poynter.indiana.edu/pait). is that the signers have taken a puting Artifacts.” We use “com­
During the workshop, we started great deal of care with the words puting artifact” for any artifact
to develop a short statement about in The Rules, so the excerpts here that includes an executing com­
the ethics of developing computer don’t constitute an official ver­ puter program. [This includes]
systems, and this statement has sion. Please visit https://edocs. software applications running on
since evolved into a document uis.edu/kmill2/www/TheRules/ a general-purpose computer, pro­
about moral responsibility. It’s not moralResponsibilityForComputer grams burned into hardware and
a Wiki, but 50 people, including ArtifactsV27.pdf for the full ver­ embedded in mechanical devices,
academics and IT professionals, sion, which includes more detailed robots, phones, webbots, toys,
have already contributed to it. An explanations for the definitions and programs distributed across
early working title was “Principles and rules. more than one machine… . We
Governing Moral Responsibil­ [include] software that’s com­
ity for Computing Artifacts,” but Preamble mercial, free, open source, recre­
most of the time, we just call it As computing artifacts become in­ ational, an academic exercise, or a
“The Rules.” creasingly complex, some have sug­ research tool.
gested that such artifacts greatly
The Rules complicate issues of responsibil­ A Working Definition of “Moral
Currently, the document includes ity. In order to help deal with these Responsibility.” We use “moral
a preamble, some definitions, five complexities, we propose five rules responsibilit y for computing
rules, and explanations, though it as a normative guide for people artifacts” to indicate that people
could change before this column who design, develop, deploy, evalu­ are answerable for their behavior
is published. Anyone can volun­ ate, or use computing artifacts. Our when they produce or use com­
teer to join the Ad Hoc Committee aim is to reaffirm the importance puting artifacts, and that their

1520-9202/11/$26.00 © 2011 IEEE Published by the IEEE Computer Society computer.org/ITPro 57


IT Ethics

actions reflect on their character.1 with other people who design, case of a crash. It seems counter-
“Moral responsibility” includes develop, deploy or knowingly intuitive to us to assign moral re­
an obligation to adhere to reason­ use the artifact as part of a so­ sponsibility to the driver for the
able standards of behavior, and ciotechnical system. use of that artifact.
to respect others who could be However, when someone know­
affected by the behavior. We do Rule 2 ingly and intentionally uses a par­
not address legal liability in this The shared responsibility of ticular computing artifact, that
document. computing artifacts is not a person takes on moral responsi­
zero-sum game. The respon­ bility attached to that use. A dra­
A Working Definition of “Socio­ sibility of an individual is not matic example is when someone
technical Systems.” Each com­ reduced simply because more launches a cruise missile at an
puting artifact should be under­s tood people become involved in de­ enemy target; a more mundane
in the context of “sociotechnical signing, developing, deploying, example is when someone searches
systems.” A socio­technical system or using the artifact. Instead, the Web for information about a
includes people, relationships a person’s responsibility in­ prospective employee. The moral
between people, other artifacts, cludes being answerable for the responsibility of a user includes
physical surroundings, customs, behaviors of the artifact and an obligation to learn enough
assumptions, procedures, and for the artifact’s effects after about the computing artifact’s
protocols.2 deployment, to the degree to effect to make an informed judg­
We acknowledge the impor­ which these effects are reason­ ment about its use for a particular
tance of sociotechnical systems to ably foreseeable by that person. application.
the issue of moral responsibility It is not our intent to absolve the
for computing artifacts. For ex­ … By using the word “foresee­ users of computing artifacts from
ample, a GPS navigator is a com­ able,” we acknowledge that the moral responsibility if they are
puting artifact, but in isolation people who design, develop, de­ willfully ignorant about artifacts
from the satellites it uses to ascer­ ploy and use artifacts cannot rea­ or their effects….
tain location, it can’t perform its sonably be expected to foresee
function…. all the effects of the artifacts, for Rule 4
[Ignoring] the sociotechnical all time. However, implicit in our People who knowingly design,
systems in which a computing use of this word is the expectation develop, deploy, or use a com­
artifact is embedded is folly, [but] that people make a good faith ef­ puting artifact can do so re­
including all relevant sociotechni­ fort to predict the uses, misuses, sponsibly only when they make
cal systems components in every and effects of the deployment; a reasonable effort to take into
discussion of moral responsibil­ and to monitor these after de­ account the sociotechnical sys­
ity involving a computing artifact ployment. Willful ignorance, or tems in which the artifact is
will make it impractical to assign cursory thought, is not sufficient embedded.
meaningful responsibility to the to meet the ethical challenges of
people most directly involved with Rules 1 and 2…. Sociotechnical systems are in­
that specific artifact. To negotiate creasingly powerful. If people
this tension, we first discuss moral Rule 3 thoughtlessly produce and adopt
responsibility for computing ar­ People who knowingly use a these systems, they are, in our
tifacts in a more focused sense particular computing artifact opinion, being morally irrespon­
(Rules 1, 2, and 3), and then place are morally responsible for that sible. Ignorance is not a justifi­
this discussion into a broader con­ use. cation for harms associated with
text that explicitly includes socio­ sociotechnical systems and the
technical systems (Rules 4 and 5). The word “knowingly” is prob­ computing artifacts imbedded
lematic in Rule 3, but we think it in those systems. Security issues
Rule 1 is, on balance, appropriate. People that occur when computing arti­
The people who design, de­ who “use” a particular computing facts are deployed via the Internet
velop, or deploy a computing artifact might not be aware of this are an example of the interaction
artifact are morally responsible use. For example, a driver might of an artifact and a sociotechnical
for that artifact, and for the fore­ not have any knowledge of a com­ system.
seeable effects of that artifact. puting artifact embedded in the Rule 4 is intended to be a pro­
This responsibility is shared car that records data for analysis in gressively heavy burden. It requires

58 IT Pro May/June 2011


an honest effort to identify and distributed system or an emerging nine different countries, united by
understand relevant systems, com­ system, the rules still hold for the their willingness to publicly assert
mensurate with one’s ability and people associated with the pieces their support for The Rules.
depth of involvement with the that are distributed, for the people

I
artifact and system. Thus, the associated with the organization think people are interested in
burden is heavier for those with of the overall system, and for the The Rules in part because the
more expertise and more influ­ people responsible for the sys­ pace of technological change
ence over the artifact’s effects and tem from which the new system and global reach of computing
over the system’s effects. Those in emerged…. and telecommunications systems
design and development cannot are unsettling. We’re hungry for
shift their burden to the users (see A New Community more clarity about who is respon­
Rule 2), and users cannot shift the On 4 March 2011, at the annual sible for what in these increasingly
burden to developers when users’ meeting of the Association for important sociotechnical systems,
local knowledge is critical to ap­ Practical and Professional Eth­ and The Rules are one attempt to
propriate ethical action…. ics, a panel of people who helped reason together about these diffi­
write The Rules discussed what cult issues.
Rule 5 the document means. I hope you’re sufficiently in­
People who design, develop, Michael Davis (Illinois Insti­ trigued to read The Rules in their
deploy, promote, or evaluate a tute of Technology) discussed entirety (available at https://edocs.
computing artifact should not both similarities and differences uis.edu/kmill2/www/TheRules),
explicitly or implicitly deceive between The Rules and codes of and I invite you to get involved. If
users about the artifact or its ethics. For example, the subject you like The Rules, you can sign
foreseeable effects, or about matters are similar, but codes of on by emailing me at miller.keith@
the sociotechnical systems in ethics are usually aimed at orga­ uis.edu. If you don’t like The Rules,
which the artifact is embedded. nizations; The Rules are aimed at you can also get involved by sug­
people from different professions. gesting changes. Perhaps the next
Morally responsible use of com­ Organizations adopt codes of eth­ version of The Rules will include
puting artifacts and sociotech­ ics, but individuals sign up for some of your ideas about our
nical systems requires reliable The Rules. moral responsibility in designing,
information about the artifacts Chuck Huff (St. Olaf College) developing, and employing com­
and systems. People who design, compared The Rules to a pro­ puting artifacts.
develop, deploy or promote a phetic voice. Although most of the
computing artifact should provide five rules are stated as explanations, References
honest, reliable, and understand­ Huff views them as inspirational, 1. M. Davis, “‘Ain’t No One Here
able information about the arti­ not just descriptive. The word But Us Social Forces’: Construct­
fact, its effects, possible misuses, “should” appears explicitly in only ing the Professional Responsibil­
and, to the extent foreseeable, one rule, but the document it­ ity of Engineers,” to be published
about the sociotechnical systems self is in the spirit of “should.” By in Science and Engineering Ethics;
in which they think the artifact signing up for The Rules, people w w w.springerlink.com /content /
will be embedded…. are embracing the responsibili­ 33338u607x251074/.
ties given there. The Rules can 2. C. Huff, “Why a Sociotechnical
Computing Artifacts that are be viewed as an attempt both to System?” ComputingCases.org,
Not Exceptions to the Rules challenge computing profession­ http://computingcases.org/general_
No matter how sophisticated als and users to embrace their re­ tools/sia/socio_tech_system.html.
computing artifacts become, the sponsibilities and to support those 3. H. Nissenbaum, “Computing and
rules still apply. For example, if an who do. Accountability,” Comm. ACM, Jan.
artifact uses a neural net, and the Ken Pimple (Poynter Institute) 1994, pp. 72–80.
designers subsequently are sur­ emphasized that a community has
prised by the artifact’s effects, the begun to form around The Rules Keith W. Miller is the Schewe Profes-
rules hold. If a computing artifact through the document’s coopera­ sor in Liberal Arts and Sciences at the
is self-modifying and eventually tive development. The community University of Illinois at Springfield. His
becomes quite different from the already contains academics and research areas are computer ethics and
original artifact, the rules still practitioners, computer scientists software testing. Contact him at miller.
hold. If a computing artifact is a and philosophers, and people from keith@uis.edu.

computer.org/ITPro  59

You might also like