Sept. 13 2006 12:43 PM

The "Peace Bug": Rethinking NATO Enlargement

By Malcolm Gladwell
       Over the next several months, the 16 members of NATO will be considering the question of if--and when--to dramatically enlarge the Western alliance. The expansion of NATO has been strongly backed by the Clinton administration, which seems intent on gathering the former nations of the Warsaw Pact into much closer contact with the United States. This, by itself, would be an admirable goal. But in all of the discussion, one critical issue has been overlooked: the impact of the sudden NATO inclusion of former hostile nations on the Pentagon's pre-programmed war-game scenarios.


       The lack of attention given this critical issue is understandable, given that the true extent of the "peace bug" did not come to light until recently. But last week, in in-camera sessions of the Senate Intelligence Committee, several former high-ranking officials testified that if Russia and Poland were to join NATO, their sudden inclusion in the intelligence database could cause the Pentagon mainframe to crash, leaving the American military without its central means of analyzing intelligence data and planning military contingencies. So far no one knows the likelihood of such a software breakdown, or how difficult it will be to avert the problem. But the fact that the Clinton administration is pushing ahead with NATO expansion anyway--given the extraordinary electronic risks it poses--suggests once again that the White House is willing to put ideological goals ahead of the national security.
 The problem at the Pentagon is not dissimilar to the so-called "2000" bug in domestic computer programs. Without costly and immediate revision of their software, hundreds of thousands of computer systems around the world will crash on the first day of the next century, because their dating systems have been pre-programmed only to accept years beginning with the number 19. In the Pentagon's case, a similar algorithm was written at the height of the Cold War in which, for war-gaming purposes, Russia and Poland were permanently encoded as enemies of the NATO alliance. (East Germany, Czechoslovakia, and several other Iron Curtain nations were, for reasons that were unclear, left out of the equation.) In the years since then, literally thousands of scenarios have been run, as the Pentagon created an enormously sophisticated "neural network" for evaluating threats to American security. No one, however, ever considered modifying that initial stretch of computer code. Now, say Pentagon experts, that binary sequence is so deeply embedded in the software that it would be all but impossible to extract without causing the entire system--and with it the military's accumulated war-gaming expertise--to crash.
       Military experts were initially confident that this problem could be finessed. Last year, for example, the Clinton administration sidestepped the issue of NATO expansion by proposing the "Partnership for Peace," a negotiated arrangement with the former Communist states in which America offered a variety of political and military arrangements while stopping short of offering full NATO membering. The Partnership was viewed at the time as a diplomatic triumph. In fact it was a clever attempt to finesse the "peace bug" problem, since the designation "peace partner" still allowed the Pentagon computers to classify Russia and Poland as enemies.
 Now, however, the Clinton administration appears to be recklessly going ahead with plans to offer all or most of the former Warsaw Pact nations full NATO status. This plan has been properly criticized because--among other reasons--of the projected expense. (According to one study, NATO expansion could cost as much as $70 billion.) What should be obvious, however, is that much of that money will not go to rebuild the new member states but rather to the programming department of the Pentagon itself, as the military establishment scrambles to replace some 30 years of accumulated wisdom in a matter of months. Surely this is a Pentagon boondoggle that puts all previous boondoggles to shame. And even if such a task is possible--and there are many inside the intelligence community who doubt that it is--there is no question that the process of rebuilding the military database will create a temporary period when the Pentagon has absolutely no idea who its enemies are. In these dangerous times, that is a risk we, as a country, cannot afford to take.

Malcolm Gladwell is a staff writer at The New Yorker.