What is strategic intelligence currently used for? First and foremost, for correctly orienting the long sequence of decision makers’ interpretations.
Secondly, for acting as automatic or non-automatic selector of relevant or non-relevant facts and news for those who have access to the intelligence system.
Finally, for correctly connecting intelligence data with the rest of open source news and the various perceptions on a topic.
Intelligence is never just a sequence of data collection. Certainly data is needed to qualify, think, imagine or refuse any Intelligence Service’s operation. What is really needed, however, is never mere data, but the indication of how the opponent secretly thinks and, therefore, what he/she selects as primary concept and, more generally, how the enemy hierarchizes and interprets his/her notions.
The Strategic Intelligence System (SIS) produces the information needed by the most important decision-makers.
Therefore, it must be simple, immediate and clear -considering that very rarely decision-makers have already experience of Intelligence Service -but also new, fresh, reasonable and, above all, capable of being even counterfactual, where needed.
If, as often happens today, even in Italy, the Intelligence Service produces models that confirm the ideas of the most superficial politicians, it is not good.
Not even for the insubstantial careers of the fools who always say yes.
In other words, an analysis that is not obvious, not always inferable from the most well-known facts, not childish and in any case not taken for granted.
Vaste programme, as De Gaulle said when he was proposed the abolition of idiots.
There is, on the one hand, the childish and very “American” fascination for new technologies, which are undoubtedly extraordinary.
Technologies which, like Artificial Intelligence, can expand, automate and make the collection and processing of intelligence data even more refined. But technologies which, each time, must be adapted to a context in which also the enemy uses AI.
Sure, but it is anyway necessary to deal with staff suitable for analysing the data sequence of an AI system and understanding how it relates to the opponent’s decision-making, whether it has to do with AI networks or not.
Either there are technology experts who understand nothing about intelligence, or there are intelligence experts who know nothing about AI technology.
What if the enemy produced – as has already happened – fragments of voluntarily manipulated information so as to later put out of phase the AI machines that interpret government’s choices from outside?
What if decision-making totally hid its operating mechanisms, thus artfully eliminating any signal capable of bringing the analytical system into its decision-making mechanism? It takes so little, indeed.
Hence we need to see how and to what extent the Artificial Intelligence subsets, cloud computing, machine learning, problem solving and robotics are really useful for intelligence operations.
In the U.S. tradition – very much linked to the “machine” myth – AI allows to automate and simplify (and here there is already a danger) data collection and, in particular, the synthesis between geospatial, Signal, HUMINT and even open source data collection.
That is all well and good, but how can we avoid the opponent knowingly “dirtying” the data sequence or developing and processing models in which the various sources contrast dangerously with each other?
Either you give contradictory news, or you put a useful source in a bad light, or you create a “narrative” that you are working for Good and Democracy, and here the Western idiots will not be able to say anything.
Probably, you shall also go back to the old traditional methods: someone who infiltrates into the enemy’s ruling classes, becomes credible and then changes the enemy’s decision-making process in our favour. Or informs us of how it really goes.
In the case of Geointelligence, AI can collect sensor data very well, often very quickly. So far, all is well, but the truth lies in the brain that evaluates, just as beauty is in the eye of the beholder.
AI is also useful in computer vision and sometimes very useful also for electronic intelligence (ELINT), especially for translation between different languages. But, certainly, this is not the whole intelligence process. We are always talking about hardware, not about conceptual software.
This is what I could call “Descartes’ complex”, i.e. the typical idea of the old Western scientific mentality that we always need to see facts and then find the automatic mechanism of a phenomenon.
This is a completely wrong criterion.
In Intelligence Services’ operations the “facts” are usually not seen, if all goes well, and never have a univocal and certain “mechanism”.
Otherwise it would not be an Intelligence Service’s operation, but a simple police action or a completely public and official operation.
The mythical rationalism of the United States and of other similar countries always tends to “automate” intelligence. Hence the more the data collection of an Intelligence Service is automated, the more predictable and useless it is. Predictable especially by the Enemy.
We always need to use “lateral thinking” and serendipity. “Lateral thinking”, based on the observations made by Maltese psychologist De Bono, uses lateral observation points to solve a problem, without using the most obvious and visible “sequential” logic accepted by everybody.
You do not dig a mine in the wrong place, but in the right one.
Nevertheless, the thinking that De Bono calls “vertical” always digs in the same place, and the human mind which, like all the other organs, does not want to work too much, is attracted by the most probable, obvious and “visible” solution, i.e. what it defines as “natural”.
Serendipity is the possibility of making accidental discoveries. Indeed, it is never by chance, but it shows the imaginative and necessary potential of those who discover a phenomenon, but who know above all how to use accidental or apparently trivial information.
It is another essential characteristic in intelligence analysis.
Furthermore, some countries think that HUMINT, i.e. intelligence from human sources, can be strengthened by AI systems that collect and select the “sources” always according to predefined patterns.
Whatever is predefined must never be used in an intelligence Service, unless there are temporary guarantees. This is the Number 1 Rule. Instead of the standardization of analysis techniques, the opposite must be done in a world where “third” countries acquire powers that were unimaginable until a few years ago.
Therefore, the predefined mechanism is a severe mistake: the “sources” are trained to avoid precisely these systems.
As was also the case at the time of the Cold War, when many Soviet undercover agents infiltrated in the Western Intelligence Services were even trained to succeed without problems in the analysis with the polygraph, the so-called “lie detector”, and also created a credible, but completely imaginary and in any case unverifiable, personal story.
They indeed used serendipity and lateral thinking. The others, with their naive positivism, let themselves be fooled.
The real problem is therefore the analysis of strategic surprise: September 11 is a case in point, but surprises can be either “widespread” or “specific”.
If you donot know how to analyse surprise, it is difficult that you can really do intelligence.
From what does strategic surprise stem anyway? From the fact that you, the victim, do not know how the strategic formula of the opponent (or friend, which is the same) is composed.
If the United States had not well understood the role played in Saudi Arabia by Prince Turki, Director General of the Saudi intelligence agency from 1973 to 2001, resigning the position only 10 days before the “9/11 attacks”, probably it would have understood that a change was taking place in relations between the Arab-Islamic world and the West.
Moreover, on a private level – which in the U.S. world is always equated to the public one – there was IBM’s near bankruptcy. It was bailed out – with difficulty – with very quick operations connected to confidential information.
Well, but this is not always the case.
Indeed, the intelligence system is not a “support” to managers’ decisions, but it is its essence, regardless of what the aforementioned managers may think.
There are new tasks and functions to be evaluated such as the greater perception State managers (except the Italian ones) have of the strategic importance of their choices.
There is also the study of global trends, a naive construction which, however, serves to outline the potential of a country’s development lines.
Moreover, in the U.S. tradition, adverse transactions have only recently been correctly reported: in the past, financial transactions, the unforeseen and clearly hostile industrial acquisitions – in short, everything in business – used to take place in the global market and therefore were fine and went very well.
There is also the adverse “line” of U.S. intelligence against the policies of central banks and large E.U. and Asian financial companies to leave the dollar area, often as quickly as possible.
This is currently a central theme of the U.S. and neighbouring countries’ counter-espionage.
Therefore, two new classes of intelligence are being developed, namely financial intelligence (FININT) and market intelligence (MARKINT).
FININT resulted from the experience gained by governments in studying some agencies in the evaluation and continuation of money laundering, tax evasion and terror financing.
But there is the danger that often completely incompetent leaders base their choices not only on classified information, but also on what they themselves believe to be the direct perception of facts.
Bravo! Let us recall the analysis made by the U.S. Intelligence Services in January 2019, when they reported to President Trump that Iran was not developing a military nuclear project, and the President told them to “go back to school” and that they were “passive and naive”.
We are coming to the “decline of truth” and the rise of what is currently called “narrative” or storytelling.
Intelligence has always defined itself as “truth to power”.
Certainly there was the neopositivist, naive and often completely silly myth of creating stable and unquestionable truths, as if the Other did not know them, thus forging stable and effective mechanisms for analysing the “enemy”, as if the enemy did not know it.
Probably something changes with non-State actors, but Western intelligence interprets these structures as if they were quasi-States. This is not the case, of course.
Rather, they are ideologically cohesive groups that present themselves as States because they represent territories, albeit only with violence.
Hence, at least for the time being, strategic intelligence will be put aside by technological development, which will largely occupy only tactical intelligence. In the very short term, there will also be the necessary training – in one way or another – of the elected politicians, which shall learn – for better or for worse – how to do it.
Then there will be the ability of the automated structures to select the malware, the distorted information, the news capable of putting its own algorithm out of phase.