Open Secrets

This story on Intellipedia and related initiatives at the New York Times has been widely covered but is fascinating none-the-less.

When the starting point is complete compartmentalization it is easy for all of us to get on the openness bandwagon. In my work day in and day out I try to help my customers understand how openness and horizontal information sharing can radically change their effectiveness, especially in the face of a networked enemy.

However, what I found most interesting about this article was how it provided balance in making the case for secrecy when secrecy is appropriate. The defense and intelligence communities will quickly gain benefits from increased openness given that they are starting from zero. But they’ll have to be careful as the benefits accrue to strike the appropriate balance between open and secret; basically they’ll need to know where to draw the line(s).

One way they may do this is by facilitating their own human networks even when the information itself must remain compartmentalized.

Information systems have difficulty with age-old concepts like “need to know” and typically have to resort to less restrictive policy and role based authorization. However, if we consider search query terms as expression-of-interest “event streams” rather than just as queries, they can be used to initiate human connections between users with similar interests who may then collaborate in a more traditional need-to-know-bound way.

One thought on the concern expressed in the article about the accuracy of consensus views developed in a wiki…

To assess the accuracy of a wiki entry should require some understanding of what consensus is represented in the entry… namely, how many people collaborated on the entry and do they represent a consensus (is it a critical mass of expertise on the topic)?

The first consideration is easy to test (although it may not always be easy to tell who may have looked at an entry and provided tacit approval without actually authoring part of the entry). The second might be readily tested by combining the query event stream idea with the named authors / reviewers of a particular wiki entry.

In other words, given a bound universe of expertise within a community, it should be relatively straight forward to identify the owners of that expertise by monitoring search event streams and other information (i.e. indexed email) and then determine whether a particular wiki entry really represents the consensus view.

Query stream harvesting and email indexing would not represent privacy issues within the intelligence community, though of course they may represent security issues of their own.

Leave a Reply

Your email address will not be published / Required fields are marked *