It is well know that rapidly changing computer technology generates significant ethical dilemma which require careful analysis. The goal of such analyses may be reactive- determining the moral quality of an action- or proactive- determining a course of action. There are several approaches and interpretations of what constitutes ethical analysis M.J. van den Hoven talks of the goal of ethical analysis as achieving a cognitive state of wide reflective equilibrium (CEPE 1997) and others talk about the goal of ethical analysis as successfully completing a procedure yielding ethical decisions.  

   The procedural approach to ethical analysis has at least two significant approaches- heuristic and algorithmic. For example a heuristic approach is taken by Bynum and Rogerson 2004 Chapter 3 “Ethical Decision Making “. It starts from a presupposition of a set of core values and takes an approach modeled on virtue ethics giving a sequence of steps calling for individual judgment based on ethical intuition.   While the Collins and Miller “Paramedic Method” is an algorithmic approach which is keyed to a rights-obligations view of ethics. The method prescribes a series of steps identifying stakeholders and their mutual obligations and rights. The analyst then follows a series of steps to identify the best alternative action in the situation under consideration. Both of these methods help to identify and integrate several elements but they optimistically presume the decision makers have a basic understanding and consensus about the nature of the ethical elements of the situation under consideration and the relative impacts of each of these elements. This problem is also exemplified in an ethical decision support process to help identify potential ethical issues in the design and development of software.

   There are two primary problems with these types of approach. First, many of the ethical questions that arise in computer ethics vary in significant ways from more traditional ethical concerns and correct results from applying these methods depends on a common interpretation of the nature of the problem.   Second, following on the first, a limited or narrow understanding of the nature of the problem contributes to a blurring of the potential ethical tradeoffs associated with a particular decision. Many of the tradeoffs are not clear when applying these methods to a complex decision. After describing this problem as it is evidenced in an analysis process a technique for reducing this problem is described and evaluated.

   Gotterbarn and Rogerson 2005) defined the Software Development Impact Statement process to help identify and mitigate the negative ethical and social impacts of software. The procedure is based in part of research which shows that software projects fail to consider all relevant stakeholders and that software risk analysis does not include a consideration of the social and ethical impacts of the software. This procedural approach to ethical analysis has four basic steps: (1) the identification of the immediate and extended stakeholders in a software project , (2) the identification of the tasks in a project, (3) for every task, the identification and recording of potential ethical issues violated by the completion of that task for each stakeholder , and   (4) the recording of the details and solutions of significant ethical issues which may be related to individual tasks and an examination of whether the current task needs to be modified or a new task created in order to address the identified concern.   (Gotterbarn and Rogerson 2005) use a method to help identify stakeholders based on Gert's moral rules [Gert 1988]. The procedure includes a set of structured questions which employs imperatives from computer codes of ethics and practice to relate the task and stakeholder. Many of the computing codes have similar imperatives. These have been reduced and categorized under general principles in the SoDIS process.

    Tests of the SoDIS procedure in industry and academe yielded bimodal results in the discovery of potential ethical risks and the identification of ways to mitigate these risks.    The first problem was that SoDIS analysts who are also computer technicians or technical specialist in other areas found it difficult to identify ethical concerns that were not immediately tied to the technical domain.   This difficulty was faced in a SoDIS analysis of the U.K. initial proposal for electronic voting.   By abstracting issues from numerous low level technical concerns analysts were able to identify a broader range of ethical concerns with the initial proposal as well as identify their potential solutions later became the basis for a series of stipulations by the Office of the deputy prime Minister (ODPM) for prospective vendors to meet when formulating the design and development of the system. (Clear, Gotterbarn, Kwan)

 

  Second, even after ethical concerns were correctly identified, some analysts 1) had problems identifying ethical risk mitigation strategies or failed to see that some strategies they proposed might resolve the identified concern but it would also 2) generate new problem. There was the further problem that there may have been alternative solutions which produced a 3) better result than the chosen solution.

   Even when all of the factual and conceptual issues are settled, there may be uncertainty about what ought to be done or decided. In this case the uncertainty of navigating the solution space generates moral problems.  Line drawing, a method developed by (Harris 2000), is a procedural technique that helps to clarify the solution space.

   Many ethical decision guidelines focus on the situation as a whole and ask about ways to respond to the situation without identifying the elements that make up the situation and which of those elements contribute negatively and positively to the situation. The failure to identify how discrete parts contribute to the situation leads to the danger that any decision may remove positive elements and increase the impact of negative elements.

     For example, both the Bynum and Collins methods ask the analyst to consider alternative actions and the consequences of those actions, but they do not provide a means to identify the significant elements in these alternative actions, an identification of which is needed to reasonably predict the consequences of the alternative actions. In complex situations there are no simple solutions but each varied element in the alternative can have multiple values; each one of which makes a significantly different solution.

     There are several advantages in using line drawing:

1. Starting with paradigm cases gives the analyst the opportunity to think about what features or attributes of an activity positively or negatively affect its ethical impact.

2. Creating a spectrum for each feature helps analysts identify dimensions to a complex problem.

3.  Analysts can take values or principles from many sources when constructing paradigm cases. Some features may be oriented around duties or rights, other features may have a benefit/harm orientation, and still others may reflect virtues.

4. Case-based analysis results in charts that match the intuition that there are ethical gray areas. Sometimes an aspect of an action may not be completely bad or completely good, but somewhere in between. The overall activity may be good in some respects, but bad in others.

   While this technique can stimulate one's thinking and lead to some important insights, it has several limitations (Harris 2004):

"First, the more ambiguous the case, the more we must know about the particular circumstances to determine whether it is morally acceptable or morally wrong."

"Second, imposing a line of demarcation between some of the cases in a series involves an element of arbitrariness."

"Third, ... concentrating on only one feature will usually be insufficient to determine where on the continuum to place a given case."

"Finally, ... the method of line-drawing resembles 'a kind of common-law ethics' in which, as in law, what one decides in one case serves as a precedent for similar cases."

  The line drawing techniques helps manage the complexity issues of software development ethical analysis of the solution space for making decisions it significantly more robust and dependable. Further work needs to be done to determine if using it will have similar benefits when used to support other analysis methods.

References:

-Bynum, T. and Rogerson, S. Computer Ethics and Professional Responsibility (pages 60-85) copyright Blackwell Publishing 2004

-Clear, T.; Gotterbarn, D.; Kwan, C. “Managing Software Requirements Risks with Software Development Impact Statements” New Zealand Journal of Applied Computing , 2006

-Collins, R. and Miller, K. “Paramedic Ethics for Computer Professionals,” Journal of Systems and Software, 17, 1 (January 1992)

-Gert, B. (1988) Morality. Oxford: Oxford University Press.

-Gotterbarn, D. and Rogerson, S. (2005a), “Responsible Risk Analysis for Software Development: Creating the software development impact statement”, Communications of the Association for Information Systems, 15 article 40 June 2005

-Gotterbarn, D. and Rogerson, S. (2005b) “Software Design Ethics for Biomedicine” S. Nagl and R. Begent (Eds.), Cancer Bioinformatics: From Cancer Biology to Therapy Design and Treatment, London, Wiley & Sons.

-Harris, C.E. Jr, Pritchard, M.S., and Rabins, M.J. Engineering Ethics: Concepts and Cases. 3rd ed. Wadsworth 2004

-Jonsen, A. R., and Toulmin, S. The Abuse of Casuistry: A History of Moral Reasoning. University of California Press, Berkeley, CA, 1988.

- (2003) “The Government's Response to the Electoral Commission's Report: The Shape of Elections to Come – A  Strategic Evaluation of the 2003 Electoral Pilot Schemes”,  

(No. Cm 5975). London: Office of the Deputy Prime Minister

- Maner, W.  “Heuristic Methods for Computer Ethics”, Metaphilosophy 33,3, 2002 pp. 339-365    Available at http://csweb.cs.bgsu.edu/maner/heuristics/maner.pdf.

-Quinn, M. J. Ethics for the Information Age Second Edition, Addison-Wesley, Reading, MA, 2006.

-Tavani, H. M. Ethics and Technology: Ethical Issues in an Age of Information and Communication Technology John Wiley and Sons, Hoboken, NJ, 2004.