Isla di Elba, Italia
At the core of the drone program is the theory of signature targeting. It is a theory based on the idea that analysts can discern “enemy combatants” from an analysis of outward indicators of patterns of behaviour.
As a practical matter, this approach requires the construction (really a primitive synthesis) of some kind of template describing the universe of behaviour indicators before the fact. In step 2, analysts can compare the observed indicators implicit in a suspected target’s individual behaviour to that predicted by the template. Step three is to make some kind of assessment — usually a subjective probabilistic calculation — of whether or not the target is a valid target. The calculation is subjective because the indicator is based on an array of assumptions buried deeply in the template.
In its more sophisticated — read pseudo scientific — incarnation, the comparison can be done mathematically on a computer using artificial intelligence algorithms. Often this effort involves statistics, including sometimes Bayesian statistics, to give the targeting “analyst” an estimate of the statistical ‘confidence’ in the result of the calculation. The mechanistic nature of this approach raises the possibility that millions of data inputs can be analyzed on a machine without human intervention, with the results summarized in a neat powerpoint briefing, and presented to an overworked decision-maker, who does not have a clue about the deeply buried subjective assumptions underpinning seemingly objective nature of the calculations.
The operative word in this application of artificial intelligence theory is “artificial” because a real intelligence appreciation, as the American strategist John Boyd showed, resides in a decision maker’s Orientation. Boyd’s theory of Orientation involves a balanced interplay of an ever-evolving synthesis growing out of many-sided, continuing analyses, which cut into a problem from many different angles. The objective is insure that one’s Orientation matches the ever-changing requirements of the changing and often menacing external environment. I worked closely with Boyd while he was developing these ideas. For readers not familiar with his ideas, my presentation, Evolutionary Epistemology, is a personal interpretation of his basic theory of knowledge generation. The goal is understand how one evolves an adaptable Orientation to support Observations and Decisions that work in changing conditions of the real world. Readers interested in learning more about Boyd’s ideas can find them explicated in Coram, Fallows, Richards, and Ozinga. Boyd’s own words (beware, they are cryptic) can be found in the Boyd Compendium.
Boyd’s theory of Orientation is crucial for understanding why an algorithmic template based artificial intelligence is so stupid, and so easy to game by all sides, for all sorts of nefarious purposes.
When viewed through the lens of Orientation, the template described above is merely a hardwired (non-evolving) Orientation based on a one-sided analysis. The fatal flaw in this conception is fundamental: Template based artificial intelligence schemes confuse analysis with synthesis. (BTW … ‘Connecting the dots’ is a popular phrase making the same mistake — which is the central reason we never seem to be able to connect the dots!)
The confusion of analysis with synthesis in template-based decision making scheme sets the decision maker up for incestuous amplification, which occurs when the preconceptions implicit in the template drive the both the analyst and decision maker to see what he/she wants to see, rather than what is. The epistemological flaw is basic: Logically, the template requires the analyst to know every possible pattern of a target’s behaviour before the fact. The unrealistic assumption of godlike omniscience reduces the subtle problem of target identification to a simple one of comparative analysis (‘observations’ vs. those predicted by the template) and leads naturally to a mechanistic statistical analysis, together (usually) with a determination of acceptable confidence levels.
In the best of worlds, this kind of hard-wired orientation can be true only if the target is too stupid to adapt unpredictably to the changing conditions that threaten its existence — i.e., if the ‘target’ is an unthinking automaton with no incentive to survive or win. But human beings are adaptive, thinking, creative, and therefore unpredictable creatures with a will to live: Already, for example, there are signs some Al Qaeda wannabees are trying to evolve counter-measures to drones — some laughingly simple and sure to be effective — to neutralize what little effectiveness drones might have. (We already know drones are killing too many civilians and producing highly counterproductive grand-strategic blowback effects).
No doubt, at this point some readers may be wondering where I am going with mumbo jumbo about the mindlessness of confusing synthesis (and creativity) with analysis (and predictability). Well, I urge you to read carefully this important McClatchy report by the inestimable Jonathan Landay. Landay explains in mind numbing detail how you and every American are now in the process of being ‘templated’ by your neighbors and coworkers. Landay shows how your friends and colleagues are being enlisted to feed indicators into the crosshairs of unseen templates concocted by anonymous, mindless, artificial intelligence apparachiks in the Federal Government and their privatized accomplices in industry, all of whom are cashing in on the self-organizing politics of money in a cultural evolutions that depends on a compliant American citizenry, anesthetized by fear.
Welcome to the Amerikan Deep State, presided over by the most superficially attractive president since John Kennedy, where signature targeting is coming home to roost in your living room, brought to you by the not-so-creative geniuses in NSA, Microsoft, DOD, DHA, Booz-Allen, etc.
Franklin “Chuck” Spinney is a former military analyst for the Pentagon and a contributor to Hopeless: Barack Obama and the Politics of Illusion, published by AK Press. He be reached at firstname.lastname@example.org