Conversation and the cult of expertise

It's a sign of my late-blooming maturity (my 56th birthday is coming around but I still dress as if I'm going to summer camp) that I agreed to participate in a conference with the CIA about how social software could help its intelligence analysts. I was brought up to believe that the CIA was one of the great evils in the world, responsible for assassinating foreign leaders and overthrowing regimes at the whim of our government and large economic institutions.

I still believe that. And now there's no question that the CIA kidnaps and tortures people. So, why would I want to help the CIA? Because good information can keep us out of unnecessary wars and can disrupt plans to kill me and my children. So, when the intelligence analysts side of the agency invited me and five other people in for open-ended discussions, I went. It was a remarkable day and a half. About 60 people from the agency were there. It was panel format except for a late afternoon of break-out sessions. Many of the attendees were already using social software such as blogs and wikis, and most were well-disposed toward it. There were some managers who were not sold on the idea, but were open-minded enough to spend time hearing about it.

The meeting began with an analyst, Todd (no last name, no e-mail address) telling us about a typical day. His primary job is preparing reports on his area of expertise. There's some informal collaboration, but at the end of the day, he gets evaluated and promoted on the basis of what he publishes. Reports frequently get boiled down to one page for our leaders, on the grounds that the more power you have, the less you need to know, a hypothesis I think we have firmly disproved in the past six years. (Look, I'm a Boston liberal. It's like having Tourette's. If it offends you, just politely pretend you didn't hear it.)

It's a reasonable system, given that you're committed to the idea of hiring experts and setting them up in their own cubicle. But it's no longer the best way to create knowledge. It draws lines that divide rather than ones that connect. It establishes an artificial "pub date" after which a document is considered to be done. Of course, the document is really only getting more and more out of date as soon as it's published.

And the CIA, typical of most knowledge organizations, draws lines around people, confining experts to their areas of expertise. This assumes knowledge is a property of experts. But better knowledge is a property of conversations. An expert's ideas and knowledge can only be improved by being put out in public (well, the closed public of those in the agency with the proper clearances) to be debated and augmented. Mistakes of fact and limitations of viewpoint inevitably emerge.

Instead of thinking that topical knowledge exists in the heads of experts, we now have the ability to go back to the original meaning of topic: topos, or place. By creating intranet places where experts can share and debate what they know, new, better and more timely knowledge emerges.

Plus, having topical places lets an organization draw the right kind of lines, the ones that connect ideas. A richly linked intellectual infrastructure creates institutional memory and enables people to get the context they need whenever they need it.

Two big issues stand in the way. First, we know how to compensate Todd for turning in good reports. We don't know how to compensate him for being a valuable member of a knowledge landscape.

Second, for security reasons the CIA assumes employees don't have a right to know something unless they have a need to know it. This is a serious restraint on sharing knowledge. The CIA's solution will be unique to its unique restrictions.

But your organization isn't the CIA and most of its security concerns it makes up just to make itself feel important. So, don't be like the CIA in this regard. (Or in the torturing regard.)

But you should be like the CIA in seriously considering how socializing knowledge can make your organization smarter.