PREDICTIVE ALGORITHMS, COVERT ACTION, AND GEOPOLITICSFebruary 24, 2014
This article came to me from a regular reader here, Mr. V.T., and I think you will see why it made the final cut, once you see what it's about:
Viewed through the lenses of High Octane Speculation, there is so much about the implications of this that disturb, one scarcely knows where exactly to begin. As most readers here know, I've advanced all along the hypothesis that the NSA's electronic eavesdropping is as much about (1) finance and (2) sociological pattern modelling(and hence, about prediction) as it is about (3) security from "terrorism". The fact that this is being done in a university research lab suggests that we are looking at the tip of an iceberg, that we are looking at research being conducted in a limited area and orchestrated by hidden hands within the black projects community, who will apply whatever results are obtained secretly, and tweak the process, if it has not already done so.
Why this would be of special concern to such black projects gurus is immediately evident: if, for example, a predictive algorithm could be extended both in its accuracy, and in the lead time of its prediction, appropriate responses can be planned, either to ameliorate or exacerbate the conditions being detected, and steering or channeling such activity accordingly.
But this in turn invokes another, equally disturbing possibility, namely, that such algorithms become, in effect, feedback loops on themselves, predicting behaviors that are the inevitable result of such covert responses, a kind of computerized and statistical version of self-fulfilling prophecies.
However, this is a still further implication, namely, that such predictive algorithms could also conceivably begin to function as the ultimate in "plausible deniability." What do I mean by this?
The idea is fairly simple. Say one wishes to conduct covert operations in destabilizing a government friendly to another government that one doesn't like(think of the Ukraine and Russia here). One could do so with some assurance of "deniability" if all the predictive algorithms said that such and such a country was "ripe for revolution." In the face of denunciations from other nations, one could produce the statistics and algorithms... the mathematics - for a scientismist age - would be convincing: "We had nothing to do with it; it is merely the inevitable consequence of a constellation of factors..." In other words, such predictive algorithms would be a virtual wet dream for the grandiose philosophies and ideologies of history such as Hegelianism or Marxism.
All of which brings us back to the sheer size and scale of that NSA facility in Utah, for the size and scale of that computing power suggest to my mind, and suggest it very strongly, that such large scale algorithmic prediction is already under way. After all, if one maintains that the real function of the spying is motivated by financial and market concerns, then this implies that the primary component of that capability is social network modeling and predictive algorithms.
See you on the flip side...