www-team  cs.cmu.edu |
Tales of Technology Pittsburgh
Post-Gazette, Sunday, Dec. 5th, 2004 Our intelligence establishment "failed to connect the dots" before 9/11. Afterward, field agents came forward with many relevant facts -- known terrorists entering the country, men taking flying lessons, etc. -- that might have predicted the attacks in New York and Washington, D.C.; but no one who could do anything put those facts together. The problem is that the government hierarchy makes intelligence information flow mostly upward and inward. Information flows from field agents to office managers, to regional managers, all the way up to the president. Hierarchies are great for command and control, but not for discovery. The problem is that two pieces of information in different parts of the hierarchy might never be combined because neither, by itself, is significant enough to be sent upward. If all information were sent upward, the people at the upper levels would be overloaded. I participated in a Markle Foundation study that proposed the use of information technology to change this system. We recommended that the agencies use an Internet-style network called SHARE to help to monitor the sharing of intelligence information. Our slogan was that the "need to share" trumps the "need to know." We hope the SHARE network breaks down the barriers; but I think more is needed. There is an old saying in the computer business: "You can lead a horse to a computer, but you can't make him type with his hooves." Whatever technology is provided to the intelligence community won't change ingrained attitudes and motivations. The people work very hard, but their orientation is not quite right. George Friedman, in his fascinating "America's Secret War," contends that our intelligence services are far too fact-oriented and not analytical or theory-oriented. There are always too many facts. In the absence of a real event such as 9/11, one needs a forceful presentation of a theory to bring just the relevant facts forth. For example, if Tom Ridge broadcast a theory that there is going to be a nuclear event in Manhattan in the next seven days, everyone who heard him would search their information for relevant facts. Of course, Ridge wouldn't do that without evidence because he doesn't want to panic anyone or look like a fool. So there's a chicken and egg problem: You need facts to justify a theory, but you can't organize the facts without having a theory. There is a motivational problem in the intelligence business. Suppose I am an analyst who has a brainstorm and see a way in which some dots can be connected. It's just a theory, but I really believe it. If I take it to my boss, he might be skeptical and demand more evidence that I can't produce. If my theory proves wrong, I will lose credibility. Better to keep quiet and collect evidence. Furthermore, rewards and status accrue to those whose ideas do rise in the hierarchy so people might unconsciously send up only that information that is likely to be accepted as fact-based. Failure of imagination is the result. There are other, nonhierarchical organizations that are good at getting at the truth -- namely, the scientific and investment communities. The completely open scientific system is probably inappropriate, but the investment system offers promise. The investment community works through markets. Lots of information is published by and about businesses, and the markets render judgment of the community. "Is IBM going to do well next quarter?" is answered by its stock price, which has been arrived at by the judgments of millions of individuals, some of whom have good theories and some of whom have critical information. This system is nonhierarchical, anonymous and rewards people with correct answers. Instead of risking his reputation on an improbable idea, an investor risks money. They are rewarded for being right. So Tom Ridge should run a futures market within the intelligence community. For example, an analyst could bet that a terrorist attack would occur in a particular city during a particular week. If it occurred he would receive a big payoff, anonymously. Ridge and his staff would watch the market and be alerted whenever a particular future spiked upward. They would heighten countermeasures, possibly preventing an attack. The analyst would lose money, but we'd all be happier. A completely public futures market was considered by the Defense Department a few years ago and was widely denounced. The notion of anyone making money from a human disaster is offensive. It would be even more offensive if the terrorists themselves made money, a most obscene case of insider trading. But their greed could tip us off and prevent the attack. The money Halliburton makes doesn't generate such useful information.
|
......