Last June, the US Air Force held an exercise codenamed Stoic Wombat, the first ever attempt to test a global intelligence apparatus in its ability to generate targets for large scale combat. If you think targeting is unimportant or just a matter of military housekeeping, think again. It is the centerpiece of American military everything — how to find something, characterize it, disable or destroy it, in the air, on land, at sea, in space and even in cyberspace. Almost the entire enterprise of military intelligence is oriented towards collecting and processing information relating to the location and nature of an object, from a base to a precise location on the battlefield, down to the level of a person, a specific building, or a network node.
Find the target, hit the target. When counter-terrorism was the American everything over two decades, that mostly meant individual members of al Qaeda and other organizations. And though that process continues in Syria and Iraq, Yemen and Somalia, as well as in various countries in Africa, the focus now is the big four: Russia, China, North Korea, and Iran. Given the global emphasis and the scale of the task means that the focus is diffuse and yet hyper specific at the same time. A multi-billion dollar machine constantly collects data — whether intercepts of signals or photographs of objects (in a variety of forms) — and the inteligence apparatus analyzes the intelligence take to pick out what is important in order to generate targets, whether that be targets for right now or the future.
I hesitate to say analyze because that suggests thought. Little thought is put into targeting in the task of penetrating the Niagara of information. Yes, tens of thousands of contractors constantly work on the software that parses the data and turns it into usable databases, but it is all too much for people. Way too much data is collected for humans to process. The speed of war ever increases. And weapons become more and more precise, requiring ever more precise answers.
As smart weapons have become the norm, and as the military shifts its attention to “great power conflict,” this task of what to attack, where to attack, when and in what order has gotten out of control. The challenge is doing it all at the “speed of war,” the military likes to say. In this regard, exercise Stoic Wombat was a monumental failure. But more important, the exercise at least focuses some attention on the process of targeting and how it dominates so much else.
It all started with the 526th Intelligence Squadron at Nellis AFB, Nevada, which provided the target base for the analysts to work from: 1,000 fake aimpoints at 27 representative locations. A constant stream of electronic signals was created to cover a classified scenario over a four-day period of the exercise: 30,000 bits and bytes of squawking radars and signals, along with 400 manufactured images to replicate a typical workload of a targeting team. At 12 other locations, all networked together via satellite, analysts chose “deliberate” (fixed and existing) and “dynamic” (emerging) targets from the intelligence. It is a federated system of analysis: an analyst in South Dakota talks to an analyst in Alabama and they create the target graphics and aimpoint analyses for, say, a war in a specific place in Europe. The resulting targeting material of answers and recommendations is then forwarded to planners and commanders. The Air Force calls exercise Stoic Wombat “a crucial step forward” in the shift from counter-terrorism to big war.
"STOIC WOMBAT was the first exercise to challenge the targeting enterprise to deliver capabilities at scope and scale of a high-end fight," said Colonel Eric Mack, 363d Intelligence, Surveillance, and Reconnaissance Wing commander. The Langley, Virginia based wing, the largest targeting organization in the military, oversaw the exercise.
Over the course of 4-days, 150 analysts across 13 geographically-separated active, reserve, and National Guard squadrons went through the motions of dynamic targeting. It was all in a controlled environment and involved an incredibly small target base, but none of it could have been done without artificial intelligence. That means everything from the automated processing of the signals being intercepted, translating them into formats and then databases and even recommendations of priorities that the “analysts” can digest. Artificial intelligence is being used to process even larger volumes of information to produce ever more cogent elements — what targets, how important, etc., etc. — all the work that used to be done by armies of analysts and targeters with maps and grease pencils.
I certainly understand the need to keep training young people to work the software and understand their checklists as they become familiar with how to do targeting. We are already close to a decade since any large sclae combat was conducted in Afghanistan and Iraq which means as people leave the military that the new one don’t get real world experience, hence the training. But watch as “artificial intelligence” grow in the military, what are the tasks that are actually pursued. The goal today is solving a problem of the military’s own making: there is so much collected with no true focus that we need machines to do the work to sort it all out. In our lifetimes, machines will do all the work. None of this makes us any smarter. And certainly none of forces us to make decisions about priorities.