Richard Wright: Intelligence Analysis is Not Done by Computers

IO Impotency, IO Sense-Making
Richard Wright
Richard Wright

Intelligence Analysis

Follow up on Robert Steele & Anonymous: Most Analysis Software Sucks — And Story of How Steele Correctly Called BSA Not Being Signed in Afghanistan

The usefulness of computer aids to intelligence analysis (“tools”) depends a good deal on what sort of ‘intelligence’ you are talking about. Intelligence is information that has been subjected to a process of research and analysis to determine its relative accuracy and relevance. When trying to determine if “analytic software” can help this process it is necessary to look at the kind of information that is being processed.

In the field of technical intelligence, i.e. SIGINT, there are a number of “tools” that are very useful. Most of these so-called tools are retrieval programs of various sorts that allow the analyst to manipulate the data in various useful ways and some of these capabilities go back over ten years ago (clustering and linking related bits if information and geographic displays using GIS). The most important unclassified technical advance impacting on analysis today is the availability of authentic data mining programs for the analyst. Data mining is NOT simple data retrieval, as many birdbrains claiming to speak for the IC appear to believe. Data mining proper uses a suite of sophisticated algorithms capably of detecting hidden patterns and trends, finding anomalies that may not be apparent, and even changing the original query structure to reflect retrieved information. Oracle has such a program based on the Oracle relational database that has been around in one form or another for at least 15 years. Data mining obviously would be effective against “big data.” The problem with all this is that these tools are designed to make research and analysis easier especially when dealing with large amounts of unevaluated information. As “anonymous” observed they cannot replace an engaged and target smart analyst.

Click on Image to Enlarge
Click on Image to Enlarge

That said such “tool” sets are much less useful to all source analysts, especially those engaged in the process of developing strategic intelligence. This type of intelligence requires much more engagement and cognitive effort on the part of the analyst. There is no software that I know of that replicate much less replace human intuition and instincts.

Final note on “Analyst Notebook” and its successors, it was designed to encourage analyst cooperation in “problem solving” and developing joint intelligence products. As is typical within the big four of the IC (CIA, DIA, NGA, and NSA) the people who thought “Analyst Notebook” was a good idea knew nothing about analysis or the analytic processes. It represented another frantic effort by people to replace dependence on individuals and shift responsibility to a group, thereby making assessment of blame harder. These are the same folks who have been looking for that will o’ the wisp, the computer that can actually perform the analytic processes that are so difficult for them to grasp. The search for a computer that produce accurate translations is also a career goal for this crowd, but appears to be based on the assumption that translation requires no interpretation or insights by the translator.

Phi Beta Iota: We say it again, citing Robert Steele – “Technology is not a substitute for thinking.” Or as Paul Strassmann has said so often, “Information technology makes bad management worse.”

See Especially:

2014 Robert Steele: Appraisal of Analytic Foundations – Email Provided, Feedback Solicited – UPDATED

Robert Steele: Why Big Data is Stillborn (for Now) + Comments from EIN Technical Council

Yoda: Exascale by 2020? No Way, Jose! Four Socko Graphics and Bottom Line Upfront — Human Brain Still a Million Times More Power Efficient

See Also:

Analysis @ Phi Beta Iota

Richard Wright @ Phi Beta Iota