Why intelligence-sharing can't always make us safer
According to the Obama administration and its critics, U.S. intelligence agencies have a problem with information-sharing. Although this critique appears to have some merit, theory and history suggest our most recent intelligence failure is of another kind. Intelligence-sharing sounds good if we imagine the happy project of dot-connecting. But the concept sounds bad, and risky, if it more resembles the game of Telephone, in which critical information is rather predictably dropped or garbled as it is passed around. The question is: How do we get good, actionable information to the decision maker in time to make a difference?
Sharing information is not a cost-free enterprise: It takes time to pass information and time for "the community" to analyze and interpret data. Intelligence succeeds not when it paints a complete picture but when it lubricates choice -- that is, when it helps key policymakers or military officers act faster or smarter than their adversaries. Former director of national intelligence Mike McConnell recognized this point in his strategic document "Vision 2015," issued in July 2008. He identified the intelligence mission as one of gaining competitive advantage, not perfect knowledge of the enemy, an approach his successor has maintained. The purpose is not to know everything -- an impossible goal in any case. The purpose is to win.
To win against a networked adversary, the intelligence community must share critical information with decision makers but not always with every element of its own community first. Assembling "puzzles" from many pieces is often necessary for planning and strategy; it takes time and the meticulous management of databases by analytical experts. But for day-to-day operations, decision makers often hold as many or more pieces than intelligence agencies do and certainly know better from moment to moment what knowledge they need to act. In terms of tactical decisions, sharing among intelligence agencies so that an "all source" product can be generated can be a form of hoarding. It can result in finished analyses that are irrelevant, unhelpful or even harmful to national security.
To understand why, consider a historical example. During the Civil War, Gen. George McClellan uncharacteristically chased down his adversary before the Battle of Antietam largely because of one soldier's intelligence coup: the discovery of a discarded copy of Confederate Gen. Robert E. Lee's Special Order 191, wrapped around some cigars. This order revealed how Lee intended to divide his forces and where he planned to go. Although McClellan bungled the chase, history records Antietam as a win for the Union. One excellent source delivered by one trusted collector motivated McClellan to act in a way that was not perfect but was more right than wrong. He didn't know everything about Lee's situation, but he knew what he needed to know to act faster than Lee had anticipated.
Yet if this instance suggests that single, timely tips can be enough, psychological research suggests that intelligence-sharing can be downright bad. Psychology professor Daniel Gilbert observed in his best-selling book, "Stumbling on Happiness," that the only thing worse than looking for a needle in a haystack is looking for a particular needle in a stack of needles. So when an intelligence establishment composed of at least 16 federal agencies, supported by a raft of state and local law enforcement agencies, mandates an obligation to share information with each other, we shouldn't be surprised when the most critical pieces are harder, not easier, for analysts to identify. This is where proximity to decisions makes a difference. Take, for example, airline ticket agents. They might not judge a father's anxiety about his son enough to stop the son from flying, but knowing this clue when the son offers up cash to fly baggage-free could trigger timely action. Holding up the delivery of the first clue until the arrival of the second cedes decision advantage to the adversary, because the decision is made at the airline counter, not back in Washington.
To win in network warfare, then, decision makers must think of themselves as collectors and analysts, too. In real-world terms, this means that ambassadors and intelligence station chiefs who know their sources are good should be able to flag a name for airlines and counselor officers without first circulating information within the intelligence community.
It is worth remembering that Gen. Joseph Hooker, a later leader of the Union Army, was the architect of the first all-source intelligence bureau, the Bureau of Military Information, but nonetheless suffered a devastating loss at Chancellorsville. Intelligence-sharing helped him plan that battle and achieve initial surprise, but the all-source analysts couldn't keep up with the wily maneuvers of Stonewall Jackson and Robert E. Lee. For that, Hooker needed a match for Jackson's cavalry, which did intelligence on the fly. He didn't have it, and he lost.
Jennifer Sims is a visiting professor at Georgetown University and a senior fellow at the Chicago Council on Global Affairs. Bob Gallucci, president of the MacArthur Foundation, served as an assistant secretary of state in the Clinton administration. Sims has consulted for the Office of the Director of National Intelligence under Dennis C. Blair and Mike McConnell.