Coming to Terms with Anticipatory Intelligence

August 13, 2019

If the U.S. intelligence community has one unifying cultural principle, it is “mission first.” But for all the benefits that principle confers, it also comes with some potentially significant downsides. These include, a tendency to act hastily — before the situation is well understood — and the ensuing waste of precious time that such haste often begets. No better example of this tendency exists than in the intelligence community’s approach to anticipatory intelligence.

What is anticipatory intelligence? It is a relatively new type of intelligence that is distinct from the “strategic intelligence” that the intelligence community has traditionally focused on. It was born from recognition that the spiking global complexity (interconnectivity and interdependence, both virtual and physical) that characterizes the post–Cold War security environment, with its proclivity to generate emergent (nonadditive or nonlinear) phenomena, is essentially new. And as such, it demands new approaches.

More precisely, this new strategic environment means that it is no longer enough for the intelligence community to just do traditional strategic intelligence: locking onto, drilling down on, and — less frequently — forecasting the future of issues once they’ve emerged. While still important, such an approach will increasingly be too late. Rather, the intelligence community should also learn to practice foresight (which is not the same as forecasting) and imagine or envision possibilities before they emerge. In other words, it should learn to anticipate.

Fundamentally, anticipatory intelligence is about the anticipation of emergence. As clear and compelling as the case for anticipatory intelligence is, it remains poorly understood. A primary reason for this goes back to the aforementioned tendency to take hasty action before understanding.

Since the 1990s, increasing complexity has been an issue that many in the intelligence community have impulsively dismissed or discounted. Their refrain echoes: “But the world has always been complex.” That’s true. However, what they fail to understand is that the closed and discrete character of the Soviet Union and the bipolar nature of the Cold War — the intelligence community’s formative experience — eclipsed much of the world’s complexity and effectively rendered America’s strategic challenge merely complicated (no, they’re not the same). Consequently, the intelligence community’s prevailing habits, processes, mindsets, etc. — as exemplified in the traditional practice of strategic intelligence — are simply incompatible with the challenges posed by the exponentially more complex post-Cold War strategic environment.

Starting in 2012 and fueled in part by performance concerns related to the then-recent Arab Uprisings — the flash-mobs that formed in Cairo were exemplars of emergence — the intelligence community rashly embarked on a veritable smorgasbord of initiatives (including the 2014 and 2019 National Intelligence Strategies) aimed at defining and clarifying anticipatory intelligence. Paradoxically, however, all that attention did not prove to be particularly helpful as a huge number of people — many with little if any exposure to complexity science — piled on to these initiatives. Indeed, as many of these people had little if any exposure to complexity science, the intelligence community’s official definition came to prioritize mythical “emerging trends,” even though trends, by definition, are established patterns and thus not emergent. In sum, these initiatives mostly amounted to a classic case of activity masquerading as progress.

So, in an attempt to change this situation and make the discussion more productive, this article proposes to reset the conversation by elucidating a simple and succinct definition of anticipatory intelligence — one that is more in keeping with the original sense of the concept while also being accessible to those not familiar with complex systems. It then distills that definition down to a few key — and completely distinct — sub-terms/concepts.

In all, by doing these two things, this article aims to help slice through the Gordian knot that is preventing a more coherent understanding of anticipatory intelligence, and enable the intelligence community to finally make real progress.

A New — More Useful — Definition

Given the above and in the interest of a fundamental place to start (or start over), the following definition of anticipatory intelligence is proposed: “The intelligence process or practice whereby potentially emergent developments stemming from the increasingly complex security environment are foreseen via the cultivation of holistic perspectives.”

Yes, it’s a mouthful. Still, this definition contains only four key sub-terms/concepts, rooted in complexity science, which ought to enable a fundamental grasp of anticipatory intelligence: emergence, complexity, foresight, and holism. Of course, there are countless more terms and concepts associated with the complexity of the new security environment; however, the aim here is to identify an essential few that can act as a launching point for a common understanding.

For the sake of consistency, the following sections will adhere to a pattern. Each section will first explain how the sub-term/concept in question tends to be used — often incorrectly — within the intelligence community. Second, the section will provide a more appropriate understanding of the term consistent with the complexity science that underpinned the original notions of what ultimately became anticipatory intelligence. Third, it will consider the sub-term/concept in the context of the Arab Uprisings, which as an emergent phenomenon exemplifies the need for anticipatory intelligence. Finally, it will explicitly call out some of the strategic intelligence analogs to the particular sub-term/concept so as to help make the distinction as clear as possible.

Emergence (and its inflected forms)

Again, the fundamental purpose of anticipatory intelligence is to anticipate emergence. That said, “emergence” is — without a doubt — one of the most overused and misused terms in the entire intelligence community. It is common to see this now ubiquitous term erroneously used as a catchall for any future occurrence or development that may come to the intelligence community’s attention. For example, the intelligence community often — incorrectly — characterizes the outcomes associated with the linear extrapolation of discrete and established trends as “emerging.” But that usage, quite simply, does not comport with the definition and use as informed by complexity science. Truly emergent issues are fundamentally new — nonlinear — behaviors that result unpredictably but not unforeseeably from micro-behaviors in highly complex (interconnected and interdependent) systems, such as the post–Cold War strategic environment. Although emergence can seemingly happen quite quickly (hence the need to anticipate), the conditions enabling it are often building for some time — just waiting for the “spark.” It is these conditions and what they are potentially “ripe” for — not the spark — that anticipatory intelligence should seek to understand.

One of the best recent global security case studies in emergence are the Arab Uprisings, a phenomenon that could not be predicted purely from a deep understanding of the discrete elements — which is an objective of strategic intelligence — that characterized North Africa and the Middle East in 2010. Nonetheless, it could have been foreseen given a broad (or holistic) understanding of the conditions (the elements and their high interconnectivity) that characterized the region at that time.

Finally, the strategic intelligence analogs of emerging phenomena are enduring issues: those that are already discretely identified (such as those on the National Intelligence Priorities Framework) or that have already emerged and achieved a degree of attention/permanence warranting the development of deep, if often narrow, knowledge.

Complexity (and its inflected forms)

Perhaps even more than emergence, “complexity” is probably the most abused term in the entire intelligence community. That is to say, it is usually used as a synonym for “complicated.” In idiomatic English that might be okay; however, in scientific terms they are not synonymous. In the broadest sense, complicated systems tend to be somewhat discrete, ordered, and linear. That is to say, they are additive (change is incremental, and they can be understood by breaking them down), they show clear cause-and-effect, they are repeatable, and they show proportionality between inputs and outputs. In a word, they are fairly predictable. In contrast, complex systems are usually open or blurry-edged, highly networked, and often behave nonlinearly — they generate emergent behaviors. They are exponential (prone to discontinuous and/or systemic change), their cause-and-effect dynamics are uncertain (even in retrospect), they do not necessarily repeat themselves (history can be a lousy guide), and outputs can be disproportionate to inputs (for which timing matters). In other words, complex systems are dynamic, volatile, and, most of all, unpredictable.

The international environment in the Middle East and North Africa from which the Arab Uprisings emerged displayed all the characteristics of a highly complex system — especially the incredible unpredictability. To put the complexity of that situation more narratively, the open and highly interconnected nature of the system enabled the fire lit in Tunisia to cascade across the region in unpredictable and surprising ways that defied precedent and that were contingent on particular conditions at a unique moment in time.

As for the strategic intelligence analogues, they are complicated (discrete, ordered, and linear) issues such as the Cold War and the Soviet Union that constituted the intelligence community’s formative experience.

Foresight (and its inflected forms)

As a matter of common practice, the intelligence community tends to — erroneously — consider almost any future-oriented estimative approach (foresight, forecasting, projection, prediction, etc.) to be anticipatory. Also, it often uses these terms interchangeably, though they are not the same. For the intelligence community’s purposes, the distinction should come from the more imaginative (which is not to say uninformed or fictitious) process, and greater uncertainty, that is associated with foresight. Foresight involves imagining how a broad set of possible conditions (trends, actors, developments, behaviors, etc.) might interact and generate emergent outcomes. Gaming or simulation is a great example of a foresight technique. In contrast, the other approaches mentioned above usually involve the extension — or extrapolation — of identified (past and current) discrete behaviors, trends, and issues of interest as trajectories — linearly, by definition — into the future.

There was no way anyone was going to have linearly projected, predicted, or forecast the Arab Uprisings — a truly emergent phenomenon. Rather, only by thinking broadly about a wide range of future conditions would one have been capable of imagining the possible emergent future that actually transpired.

The strategic intelligence analogs to foresight, then, are forecasting, prediction, and projection.

Holism (and its inflected forms)

The intelligence community talks about analysis. After all, that is what it does — it attacks problems by breaking them down, which is what analysis actually means. And it is not at all surprising that the intelligence community does this, as linear/complicated problems, like the Soviet Union/Cold War, are well suited to reductionist approaches. However, one of the most challenging aspects of nonlinear/complex problems is that they cannot be well understood via such reductionist — that is, analytical — means. Rather, they need to be looked at more holistically — synthetically or creatively, so to speak. Such an approach is, of course, contrary to how the intelligence community has traditionally thought about problems. For instance, not only has the intelligence community tended to carve issues up into discrete and narrow organizational units, but it has also tended to block out or even ignore the influence of the United States (“Blue”) on the larger system (which is ironic, if the intelligence community truly believes the United States is a consequential actor). To reliably foster such holistic perspectives, the intelligence community needs to, among other things: educate, recruit, and retain synthetic thinkers; flatten its organizational structures; and increase use of synthetic modeling, gaming, and simulation techniques. Additionally, it must explore how artificial intelligence, machine learning, and big data solutions might prove useful in this particular context.

In the case of the Arab Uprisings these tendencies came to bite the intelligence community, as there was simply no way to understand — never mind anticipate — the larger phenomena with even the deepest knowledge of the separate components (for example, Tunisia or Egypt). Understanding would have required a holistic perspective that went against the very grain of the intelligence community’s being — cognitively, organizationally, etc.

Finally, the strategic intelligence analogues to this are, of course, notions such as analysis and reductionism that emphasize a deep but narrow understanding of the identified elements of the system and are not particularly concerned with the broad array of dynamic relationships that tie or blend them all together.

Finding the Plot

Given the above, it is vital that the intelligence community get onto the proverbial “same page” regarding anticipatory intelligence. Exponentially increasing global complexity is the defining characteristic of the age. For proof, one need only look at how the information technology revolution (the “cyber” realm) is now unrelentingly blending and merging almost every other traditional domain, discipline, and/or organizational category. However, over the course of anticipatory intelligence’s gestation, so many of the arguments over what it “ought to be” have seemingly lost the plot. Complexity science and its associated lexicon and concepts have been dismissed as jargon or, worse, employed in ways inconsistent with their real meaning. The result is that a vital capability demanded by a new era has been effectively lost in the muddle and seemingly stillborn.

This should not continue. Again, the potential — even likelihood — for emergent phenomena that will sweep across all the old domain boundaries (political, social, technological, financial, etc.) is increasing. Occurrences like the Arab Uprisings are not going to be “once in a lifetime” events. Moreover, issues that the intelligence community traditionally thought of as discrete (such as nation-states like Russia and China) no longer are. All told, new, truly anticipatory approaches are urgently required. And consequently, the intelligence community no longer has the luxury (not that it really ever did) to keep batting around notions of anticipatory intelligence in the hope of one day arriving at a common understanding.

To end (or, more hopefully, to begin again), today’s complex world is indeed different from the complicated Cold War world that constituted the intelligence community’s formative experience. Such a fundamentally different challenge demands a fundamentally different kind of intelligence. Thirty years beyond the Cold War, there is no more time to waste.

 

Josh Kerbel is a member of the research faculty at the National Intelligence University (NIU)—the U.S. intelligence community’s sole accredited and degree-granting educational/research institution. Although this article is adapted from work done at NIU and has been approved for public release, the views expressed herein are the author’s alone and are not necessarily shared by NIU, the U.S. government, or any other component thereof.

Image: Wikipedia Commons