CIA ‘Grassroots’ Campaign Boosted Intellipedia, ‘Evangelist’ Says

Rapid-fire policy decisions are “increasing exponentially” at the CIA due to the need for quick turnarounds on enemy location and activities, said Innovation Officer Calvin Andrus. His lobbying to agency brass on complexity theory, drawn from studies of how environments such as the rainforest function “naturally,” led to the creation of Intellipedia. Decisions that took up to six weeks are now made in 15 minutes, he said: “The ‘yesterday’s news tomorrow’ paradigm … is not okay now.” As with the blogosphere, the challenge has been distilling the “really brilliant ideas” posted daily from the “nonsense,” Andrus said.

“We all thought [Andrus] was crazy, talking about birds and rainforests,” said Sean Dennehy, the CIA’s Intellipedia and Enterprise 2.0 Evangelist. But Dennehy’s review of Wikipedia’s discussion pages reminded him of daily intelligence discussions. “In an alternate universe … we don’t have to pull people offline. It’s two mouse clicks away,” he said. Intellipedia, launched in 2006, actually followed a line of inventive tools that came out of Intelink, the community’s classified intranet system, Dennehy said: “It’s been a struggle. It’s a lot of cultural issues we have to encounter in bringing this kind of open-source ethos.”

The first mistake was trying to require Intellipedia use by all through top-down order of the Director of National Intelligence, Dennehy said. In the first attempted use of the wiki, to write a national intelligence estimate, “everyone around the table was looking for excuses” to not participate, and some even scoured the building looking for agency lawyers to get Intellipedia shut down. Telling staff in a “need to know” culture to use a wiki is like “telling people their parents raised them wrong,” Dennehy said. The cold shoulder given to Intellipedia, though, led to a “grassroots” persuasion effort, analyst by analyst, that has enshrined principles of openness and sharing in Intellipedia by starting with a small user base, he said.

For $2,500 upfront, the Army created CAVNET, “nothing more than a messaging board” for sharing combat data, said Lt. Col. Patrick Michaelis, who led the effort. An officer’s “brain dump” e-mail to Michaelis, pointing to time patterns in improvised explosive device attacks in Afghanistan, “reframed my entire idea of how to fight.” Army convoys started varying distance and speed between vehicles to avoid the attacks, and the online forum followed, he said. Another $2.5 million was spent to connect military users across Baghdad, and the system now receives 100,000 visits monthly.

Michaelis went to the Defense Advanced Research Projects Agency after the first leg of Operation Iraqi Freedom to create the TIGR system. Its “almost Google-like functionality” helps officers to “sift, slice and dice through massive amounts of data” such as enemy attack types and times, he said: “The troops love it.” The system’s next use will be for “nonkinetic” data such as cultural information about the local population and economic drivers, as the Army transitions to “stability operations” in Iraq, Michaelis said. But “in essence we're still culturally a hierarchy when it comes to transferring knowledge,” weeks behind an enemy that makes decisions in days.

The assumption that young analysts are the drivers of Intellipedia is ahead of actual recruitment, Dennehy said. The so-called millennial generation “hasn’t hit our workforce yet,” and many of the most active editors on Intellipedia have a “fair amount of seniority.” Dennehy’s goal has been to turn agencies into a unified “community,” in which analysts moving among the intelligence agencies can retain all their online contributions, he said. Most of the Army is under 28 and they've grown up online, Michaelis said, but there’s “hesitation” among older officers to use Web 2.0 tools. It’s not just “antiquated” policies -- the “political repercussions of a post gone bad” unnerve older officers, he said. But there’s an increasing realization of the value of “collaboration as a form of command.” Much of the change in attitude is coming from a losing public-affairs battle with the enemy, Michaelis said: “We're getting beat because we're not telling our own story.”

The risks from mobile data delivery aren’t so much about interception of classified data, but the “last-10-inches problem” of the uncontrolled environment in which data are received, Andrus told a questioner from CTIA. There are “guards and locks” protecting any person using a secure, classified network on a desktop, but classified delivery to a BlackBerry could be seen by an officer without the requisite security clearance. Michaelis pointed to a “policy-driven digital divide” created by rules on what can’t be displayed in different environments, “and sometimes it doesn’t quite meet the common-sense test.” Data sent to a digital display in a Humvee, intended for a commander with the requisite clearance, can be seen by everyone else in the vehicle, and “they better look at it” in case the commander gets attacked, he said. “There are some policy decisions that need to be re-addressed.”