Computer system reliability and nuclear war

Reviewer: Dennis Roy Thompson

At the onset, the author of this paper states, “Portions of a preliminary version of this article were presented at the Fourth Congress of the International Physicians for the Prevention of Nuclear War. . . ” This information is footnoted on the first page of this highly biased paper. Borning's prime question is, “How dependent should society be on computer systems and computer decision making__?__” especially “. . . in the use of computers in command and control systems for nuclear weapons. . .” He concludes that “The fundamental problems are due to untestability, limits of human decision making during high tension and crisis, and our inability to think through all the things that might happen in a complex and unfamiliar situation. We must recognize the limits of technology. The threat of nuclear war is a political problem, and it is in the political, human realm that solutions must be sought.” Everything in between the beginning and conclusion is a biased discussion of many important issues. Borning's references (which run to a total of 142) are also biased, for I see no poignant countervailing references like [1]. Some of the arguments he presents leave off other possibilities: “A continuing trend in the arms race has been the deployment of missiles with greater and greater accuracies. This trend is creating increasing pressure to consider launch-on-warning strategy. . . .” He does not mention that with increasing accuracy we need smaller and smaller warheads that may eventually use nonnuclear warheads. Borning presents the “false alert” of June 3, 1980, when the Strategic Air Command mistakenly reported a Soviet missile coming toward the US, and continues: “It is quite possible that some of these preparations were observed by the Soviet Union.” There is no factual evidence for this, just a presumed possibility, yet he argues that “World War I following the assassination of the archduke in Sarajevo, in which the alerts and mobilizations of the European powers interacted in just this way, is a historical precedent for this possibility. . . .” Even if World War I can be ascribed to this kind of chain reaction, this is shoddy, dishonest reasoning. If one is a slave to history, one is bound to repeat it. SDI is represented as totally unworkable owing to its complexity, both in hardware and software, which echoes the author's conclusion that we must recognize the limits of technology. We cannot have automatic systems because of the reliability problems, conceptual problems, and strategic problems. We cannot have reliable computerized systems because of human operator problems. And we cannot have humans making decisions under conditions of stress. What then is the answer if neither machines nor men (whether they be generals, privates, or politicians) can function__?__ I posit the answer—we must not recognize the limits of either technology or man. Though this is an extremely biased paper and I wonder if it should not have appeared in some political journal, it should be read; it does contain a detailed discussion of many crucial issues. The following are the major section headings: False Alerts, Tightly Coupled Nuclear Forces, Launch-on-Warning, The Reliability of Complex Systems, Limited or Protracted Nuclear War, Future Computer-Controlled Military Systems, and Conclusions.