Much has been made of the difficulties facing computer professionals who would be ethical. We have little evidence of success stories. This is an early report of a project to identify moral exemplars in computing and to use life-story interviews (McAdams, et al. 2001) to document their developmental histories and values.
In November, 2002, a panel of seven experts in computer ethics met to identify criteria for selection. Final criteria were similar to those used by Colby and Damon (1992):
- Either a) a sustained commitment to moral ideals or ethical principles in computing that include a generalized respect for humanity, or b) sustained evidence of moral virtue in the practice of computing.
- A disposition to make computing decisions in accord with one’s moral ideals or ethical principles, implying also a consistency between one’s actions and intentions and between the means and ends of one’s actions.
- A willingness to risk one’s self-interest for the sake of one’s moral values.
- A tendency to be inspiring to other computing professionals and thereby to move them to moral action.
Panel members nominated exemplars from the UK, Denmark, The Netherlands, Norway, and Sweden. The panel approved all nominees, based on biographies provided them. Concern about any nominee from any panel member was cause for removal of the name. Only one was removed for this reason. Sixty three exemplars were nominated and 35 were contacted. We conducted interviews with 24, 12 in the UK and 12 in the other countries, a response rate of 71.43%.
Thirteen exemplars had experience in academia, 15 in industry, and 3 in government. These categories overlapped, with seven exemplars having experience in two areas. Eleven exemplars were in the final decade of their careers and 4 were retired. Four were in the first decade of their career and 5 in the middle of their careers. There were 9 women in the sample.
We were able to identify exemplars in academia and industry in about equal numbers, suggesting that situational pressure in each sector does not drive out or constrict those who want to be principled in their profession.
The self-presentation of these exemplars in computing parallels that seen in much of the social service exemplar literature (see Colby & Damon, 1992). Particularly, they appear to have two orientations to their work. In the social service exemplar literature, the two approaches are called helper vs. reformer. We have called the helper orientation the “craftsperson” approach because of the ways these exemplars integrate their ethical values into their technical craft. We found more “pure” craftspeople than “pure” reformers in our sample, though many exemplars were mixtures of the two.
Craftspeople tended to:
- Focus on users: They focused on helping users/customers.
- Viewed users/customers as having needs: Their primary focus was designing computing systems to address these needs.
- Viewed barriers as inert obstacles: They viewed difficulties as puzzles to be solved, or negotiations to be worked out.
- Believe in the efficacy of helping users: They believed their efforts helped users and customers and kept their focus on that work.
- Be positive in emotional tone: Because they felt they were effective in designing technology to help solve problems for others, they were the most positive in emotional tone among the exemplars.
- Design computing technology towards ethical ends: Craftspeople saw their computing work as their way to address the needs of users or customers. Their ethical principles (e.g. user focus, customer need, software quality) were deeply intertwined with the way they approached their work of design.
Reformers tended to:
- Focus on the social system: They saw the social system as needing basic reform. This usually meant the computing industry, though some felt societal change was required for industry change.
- View individuals as victims: The people they were trying to help were seen as people who needed justice or some other public good. Thus they saw the system as lacking the values they wanted to bring to it.
- View barriers as active opposition. They viewed barriers to reform as active resistance to reform by those who had different interests.
- Took the role of moral crusader: Their approach was that of the crusader out to reform the system.
- Believed in the necessity of systemic reform: They did not think individual help to users was effective. They saw systemic reform as necessary.
- Be negative in emotional tone: Partly because systemic reform is difficult, and partly because they saw themselves as a minority trying to inculcate values in a resistant system, these often felt less effective in their efforts and had a more negative emotional tone about their work.
Most of the exemplars actively cultivated a social support network (see also Colby & Damon, 1992; Fischman, et al., 2004). Almost all of the exemplars cited a large number of people who influenced their ethical development. Those who did not were the most negative in emotional tone.
There were several reasonably pure reform exemplars. These saw their reform efforts as independent of the technical work they did on computing and saw the computing industry as resistant and inhospitable to the values they were championing. Craftspeople, on the other hand, held values already intrinsic to computing and their skills were intertwined with their technical expertise. Values such as user focus, customer need, and software quality mean specific things in the computing community. And though the implications of these values are the subject of debate, there is general consent that the values are important. This is not surprising given that value commitments are often learned as a part of technical expertise (Collins & Pinch, 2002).
There is, then, a variety of ways that computing professionals can be ethical. This finding undermines the monolithic approach and suggests that there might be a social ecology of computing ethics. In such an ecology, different skills sets, values, and emphases would work together (and occasionally at cross purposes) to support ethical concern in computing.