IN SEARCH OF A COMMON RATIONALE FOR COMPUTER ETHICS by Robert N. Barger, Ph.D. University of Notre Dame Read at the Third Annual Computer Ethics Institute Conference The Brookings Institution Washington, DC April 28, 1994 Copyright (C) 1994 by Robert N. Barger That there might someday be agreement on a common rationale for computer ethics is probably a "useful myth." It is useful in the sense that if it were not attempted we would not even come close to achieving a common rationale. Yet it may be mythical in the sense that agreement upon such a rationale is probably beyond achievement. At last year's conference, I suggested that computer ethics can be grounded in one of four basic philosophies: Idealism, Realism, Pragmatism, or Existentialism (a recent movement - Logical Analysis - claims that metaphysics cannot be discussed, but its epistemology makes clear that it brings in a Realist metaphysics through the back door). Idealists believe that reality is ideas and that ethics therefore involves conforming to ideals. Realists believe that reality is nature and that ethics therefore involves acting according to what is natural. Pragmatists believe that reality is not fixed but rather is in process and that ethics therefore is practical (that is, concerned with what will produce socially-desirable results). Existentialists believe reality is self-defined and that ethics therefore is individual (that is, concerned chiefly with forming one's own conscience). Idealism and Realism can be considered absolutist philosophies because they are based on something fixed (that is, ideas or nature, respectively). Pragmatism and Existentialism can be considered relativist philosophies because they are based or something relational (that is, society or the individual, respectively). A moment's reflection will show that if a common rationale is to be found, it will paradoxically have to have an absolutist basis rather than a relativist one. The reason for this is that, in the long run, individual or social ethical policy must be based on principle...and principle can only be grounded in an absolutist framework. Relativist philosophies (especially Pragmatism) may be helpful in the area of ethical methodology, but they provide no absolutes which can serve as a basis for ethical decision-making. Even assuming agreement on an absolutist basis for computer ethics, two further problems arise. The first is what to prohibit. The second is how to enforce such a prohibition. Rabbi Gamaliel long ago paralleled Jesus' statement of the Golden Rule by saying: "Don't do to others what you would not want them to do to you." The trouble is that, in the rapidly developing world of computer technology, we don't know what to prohibit. William Witsman, a colleague at Eastern Illinois University, recently put it this way: "We don't know what not to do." I believe his sense of the situation is correct...if not his grammatical phrasing of it. The following tale may illustrate a way of dealing with this ignorance: A priest once gave a fairly abstract sermon on miracles and was stopped after Mass by a young man who said he still did not quite understand what a miracle was. Somewhat perturbed, the priest asked the man to turn around. Then he kicked the man solidly in the seat of the pants and said: "Did you feel that, my son?" On receiving a quite affirmative reply, the priest said: "Well, if you hadn't...now that would be a miracle!" The need today, then, is to come to a decision about the things that should not be done in computing. This decision should be based on clear commonly accepted principles. In regard to enforcement, I am reminded of the approach adopted by the Rule of St. Benedict which was developed in early Christianity to regulate the behavior of monks. Benedict recommended a three-stage process for dealing with recalcitrant monks. First, he said, the errant monk should be privately admonished by the abbot of the monastery. If that didn't work, then the monk should be brought before the full monastic chapter and publicly warned. If this still had no effect, then Benedict said, "Let two of the stouter monks explain the matter to him." What we need today, after we know what to prohibit, is efficient enforcement of these prohibitions. Louis Berzai, a colleague at the University of Notre Dame, has pointed out an example of inefficient enforcement in the sentencing of Robert Tappan Morris, Jr., who infected the Internet with a worm causing untold thousands of dollars in losses. He received a fine of only $10,000 and 400 hours of community service. I join Louis in wishing, for purposes of deterrence in this precedent- setting case, that the judge would have more forcefully "explained the matter to him" by means of a stricter sentence. In summary, my suggestion for achieving a common rationale for computer ethics would be: first, an agreement on principles (e.g., honesty, justice, truthfulness), second, an application of these principles through the prohibition of unethical (i.e., unprincipled) behavior, and third, enforcement of these prohibitions by providing punishment for non-compliance and/or by providing positive incentives for compliance. As a final note, I would agree with the remarks of Professor Alicia Juarerro made earlier in this conference that we are in the middle of a technical revolution and do not yet know what is going to come out of it. But that does not mean that traditional absolutist ethics is obsolete. Consider that not only were names for inventions during the industrial and electronic revolutions drawn from the preceding technology (e.g., the "horseless carriage," the "wireless," the "electric lantern"), but mechanical engineers one hundred years later are still evaluating engines on the basis of horse power, and electrical engineers are still evaluating illumination on the basis of candle power. The former technology has been superseded, but its basic standards are not thereby made obsolete. The same may be true for the moral standards to be applied to the current technological revolution.