It is September 2025. As tensions between the United States and Iran reach a crisis, a U.S. Navy carrier and two AEGIS-class destroyers are escorting Saudi oil tankers through the Hormuz Straits. Suddenly a dozen remote-controlled Iranian Revolutionary Guard Corps speedboats roar into view, each armed with 2,000 kgs of deadly explosives, and begin wildly weaving around and circling the American escort vessels. But before they close in and detonate their charges, each of the IRGC speedboats goes up in an explosion of steel and sea foam.
How? Scores of small self-guided sea mines, each less than a foot long, had been released by a nearby U.S. submarine to track the movements of the Iranian vessels. Using their artificial-intelligence-enhanced sensors, these seagoing drones were able to predict the direction and speed of each boat’s evasive maneuvers. And when the Iranian craft moved in too close for comfort, the drones swarmed, attached themselves to the hulls of the Iranian ships, and, at the push of a button by their operator on board the submarine, blew up the boats almost simultaneously.
Maddened by failure, the Revolutionary Guard commander orders his battery of the latest Chinese-made supersonic anti-ship missiles to fire on the American ships. But U.S. Air Force reconnaissance drones flying overhead, too small to be detected by Iranian radar, intercept the order as it’s given and signal the twin battery of electromagnetic rail guns mounted on the lead AEGIS destroyer. The guns immediately pivot and rapid-fire hypersonic projectiles at a speed of Mach 7, which destroy each anti-ship missile almost before it leaves its launch vehicle.
Then, in order to deal with the missile threat once and for all, the Navy’s latest directed-energy weapon (deployed on the second destroyer) fires a high-power microwave burst that burns out all the electronics of the missile battery while leaving its Iran crew scared and bewildered but completely unharmed. Indeed, since the IRGC boats were all unmanned, not a single person has been killed. Yet America has won a major victory for control of the Hormuz Straits, as the oil tankers sail safely into open sea.
Welcome to the Pentagon’s New Look, and the future of combat in the high-tech age.
This is no Stars Wars-style fantasy. Both the Navy and the Air Force are actively working on swarm technology involving dozens or even hundreds of drones that interact with and follow their “intelligent” lead drone like a flock of birds or a school of fish, and which General Ellen Pawlikowski, head of the Air Force’s Material Command, has said “can be very much a game-changing reality for our Air Force in the future.” The Navy has been fielding its own self-guiding CAPTOR mines for several years, which are anchored on the ocean floor and fire a torpedo when the pre-programmed designated target comes within a five-mile radius. The next generation of sea mines after CAPTOR will be smaller, smarter, and able to hunt in packs—while also costing a fraction of the CAPTOR.
At the same time, the Navy aims to test its rail-gun1 prototype on a naval vessel in 2017 and deploy the weapon on a Zumwalt-class DG1000 destroyer in 2018. Two defense companies, General Atomics and BAE Systems, are locked in a head-to-head competition for the final contract. In less than a decade, the rail gun should be as ubiquitous on Navy vessels as the conventional five-inch gun is today.
As for directed energy weaponry, the Air Force and General Atomics are experimenting with a high-energy laser mounted on an existing unmanned drone, the MQ9 Reaper. General Paul Selva, the vice chairman of the Joint Chiefs of Staff, has announced that the JCS will be designating the high-energy microwave realm an official military “domain” comparable to the land, sea, space, and cyber domains. “After a nearly half-century quest,” a recent report from the Center for New American Security declared, “the U.S. military today is on the cusp of finally fielding operationally relevant directed-energy weapons”—which will eventually be as available to the individual soldier and Marine as his (or her) M-16 rifle.
All this is part of a new overall Pentagon strategy to marry the latest technologies with war-fighting. It has been dubbed, with the infelicity common to such military plans, the Third Offset. Not many people outside military circles have heard of the Third Offset, but it’s by far the best thing to come out of the Obama Pentagon—and will go a long way to determine how the United States confronts its future foes, not just Iran but Russia and China.
The Third Offset is also a way to defend America and project American power militarily in the face of flat or declining defense budgets. Its aim is to break the pernicious cycle of defense acquisition, in which the more the Defense Department spends, the less it actually gets—in terms of equipment and personnel and overall effectiveness.
That cycle veered into the danger zone during the period of the Iraq and Afghanistan wars, when military spending soared past half a trillion dollars a year (even excluding those wars) at the same time that military readiness was getting harder to maintain and the development of next-generation fighters, bombers, and warships was becoming too expensive to contemplate. (Our current state-of-the-art F-35 Joint Strike Fighter, with its $100 million apiece price tag, was first developed in the 1990s.) In the meantime, future potential opponents, such as Russia and China, were busy building, upgrading, and thinking about how to win a full-scale war with the United States.
As the name implies, the Third Offset isn’t the first time the Pentagon has confronted this problem of dealing with potential opponents under the constraint of a shrinking budget and smaller forces—in today’s case, with an Army, Navy, and Air Force that haven’t been this small since before World War II. The first effort came in the 1950s, when the Pentagon shifted to relying on nuclear deterrence to halt the Soviet surge in the early days of the Cold War. The second occurred in the 1970s with the advent of “smart bombs” and Stealth, as well as the first extensive use of computer networks for command and control, to overmatch a Soviet adversary who had added to his formidable land forces a massive nuclear-missile threat and a growing blue-water navy.
The Third Offset, however, may be the most critical of all. Today we face not one powerful and well-armed antagonist, as in the Cold War, but two: China and Russia. In the wake of President Obama’s strategic retreat around the globe, these two nations have advanced their military presence with bold moves, including Russia’s invasion of Ukraine and intervention in the civil war in Syria, and China’s planned construction of a naval base at Mogadishu in the horn of Africa—as well as military airfields on artificial islands in the South China Sea.
The goal of the Third Offset is to develop a series of “technological overmatches” against Russia and China by turning America’s two biggest economic assets, its scientific research and high-tech industries, into tools for transforming how we will fight any future war.
These run the gamut from rail guns and directed-energy weapons to cyberwar, robotics, artificial intelligence, and even 3-D manufacturing and Big Data. The last two have become increasingly common in the business world and the commercial sector, and they are destined to become important tools in America’s high-tech arsenal.
Above all, the Third Offset means a new generation of weapons that are faster, smaller, and more intelligent or “autonomous,” meaning able to perform more and more functions on their own without human guidance or control. The hottest word in today’s Pentagon is “autonomy,” and the core of the Third Offset strategy is relying on unmanned systems that spring from the Predator and Reaper drones used to hunt terrorists in places like Afghanistan and Pakistan. These unmanned systems will be considerably more sophisticated, considerably smaller, and far more numerous than their predecessors. And they will be less reliant on a command and logistical chain manned by humans, who account for one of the largest and fastest-growing costs at the Pentagon.
For some, these changes raise fears of a Terminator-style apocalypse, with “The Rise of the Machines” that rule the battlefield signaling the doom of humanity itself—while what can go wrong in a war run by machines, including in space, has been brilliantly showcased in the P.W. Singer–August Cole novel Ghost Fleet, which has become virtually required reading in today’s Pentagon.2
Many of these fears are overstated. But there are some clear dangers and risks that come with the “intelligent” battlefield. Strategists and politicians will have to reckon with them as the turn to The Fast, The Small, and The Intelligent in weaponry develops its own momentum.
All the same, the Third Offset strategy offers a way for the next president to boost American military strength and project power more effectively, within the constraints of tighter Pentagon budgets. He or she can also take full advantage of the fact that most of the technologies involved in the Third Offset—drones, lasers, sensors, robotic components, and the algorithms that power them—will become cheaper, even dramatically cheaper, over time. Moore’s Law hovers over the Pentagon’s new game plan like a benevolent protective drone, ready to drop costs even as capabilities blossom and multiply. Backed by developments in the cyber domain like Big Data, and in artificial intelligence, the new era of autonomous weapons will make the Nimitz-class carrier, the F-35, and the M-16 look like the equivalent of the musket, the mounted knight, and Old Ironsides—and at a fraction of the cost.
The bottom line is, the technologies being unveiled under the Third Offset strategy may be the last best chance we have to defend the United States when we face not one but two large and aggressive potential foes, in an era of ever more costly conventional arms and ever smaller defense budgets—as well as the best opportunity to help America’s allies do more to defend themselves by adopting the same technologies.
So how does the Third Offset work?
An offset strategy uses a strength in one area to compensate or offset a disadvantage in another. In the military sphere, this involves using a series of qualitative technological advantages to offset quantitative disadvantages on the battlefield.
History tells the tale. In the Pentagon’s view, the First Offset came when, faced with a massive Soviet advantage in ground forces, President Dwight Eisenhower opted to move forward with a nuclear-deterrence strategy that combined strategic targeting of the Soviet Union with the possibility of tactical nuclear strikes directly on the battlefield. Dubbed the New Look, this approach meant a massive reduction in manpower as well as defense spending by leveraging the advantage the U.S. enjoyed both in the size of its nuclear arsenal and its air power (signaled by the birth of Strategic Air Command), sea power (the advent of nuclear-powered and then nuclear-armed submarines), and ballistic-missile development—developments that also made it possible for America to put men on the moon.
The Soviets did not sit still. By the early 1970s, they had built up their own tactical and strategic nuclear arsenal while continuing to maintain their huge ground forces in Eastern Europe and creating a blue-water Soviet navy, equipped with their own “boomers” or nuclear submarines.
Something had to be done to avoid a massive escalation of the size and scope of U.S. military power—which, in the wake of Vietnam, wasn’t politically possible. So in the summer of 1973, the Pentagon launched a project to develop a new generation of weapons intended to respond to a Warsaw Pact attack.
Out of that initial study, and follow-up work during the Carter and Reagan administrations, came the Second Offset—as represented by a cluster of new military technologies such as cruise missiles, precision-guided munitions or PGM’s, and Stealth. As early as 1984, even before the advent of strategic missile defense (SDI, or “Star Wars”), all these forced the Soviet military to reconsider its entire formula for winning the Cold War. It also eventually spawned the high-tech, computer-networked, stealthy, and GPS-guided U.S. military that would score devastating victories in the First Gulf War, Afghanistan, and Iraq.
But “just as with the first offset strategy, the second offset strategy is showing its [age].” Those were the words of Deputy Defense Secretary Bob Work in a January 2015 speech that laid out in detail the concept of the Third Offset strategy then–Defense Secretary Chuck Hagel had unveiled the previous November.
“Potential adversaries have been modernizing their militaries, developing and proliferating disruptive capabilities across the spectrum of conflict,” Hagel had said. While America had spent the decade after 9/11 focused on fighting the war on terrorism and counterinsurgencies in Afghanistan and Iraq, Russia and China had been steadily upgrading their militaries with a host of technologies aimed at destroying or at least crippling the high-tech advantages our military had enjoyed following the Second Offset. Such technologies include ballistic and high-speed anti-ship missiles, highly sophisticated anti-aircraft missile systems, electronic jamming and cyberattack, and anti-space weapons that can kill GPS and communication satellites that the U.S. military needs to keep track of its forces and to guide its precision “smart” weapons. In addition, both the Russians and Chinese have built their own versions of Stealth aircraft. This was, Hagel declared, a “clear and growing challenge to our military power.”
Hagel’s November 2014 speech was short on details. In it he mentioned only four such “offset” technologies, three of which aren’t normally associated with military weapons: miniaturization, Big Data, advanced manufacturing, and robotics or “autonomous systems.” Work’s follow-up speech gave a more detailed shopping list that included advanced sensors and communications; “missile defense and cyber capabilities.” He also mentioned “unmanned undersea vehicles; advanced sea mines; high-speed weapons; advanced aeronautics from new engines to new, different types of prototypes; electromagnetic rail guns; and high-energy lasers.”
The new strategy isn’t just a plan to build more weapons, or even to build smarter weapons. It’s to pick a competitive advantage our adversaries can’t duplicate and, therefore, to create uncertainty in their minds about the effectiveness of their war-fighting strategy. In short, the goal is as much psychological as it is practical: to force a potential adversary to go back to the drawing board, and invoke doubt in the adversary’s own mind about his ability to meet us on equal terms on the battlefield.
If cyber technology seems the most obvious example of where the military has come to rely on developments in the civilian sector to upgrade, extend, and protect its command-and-control networks—where companies like Apple, Microsoft, and Google take the lead and the Pentagon necessarily follows—a less obvious case is miniaturization. It’s become one of the fine arts behind smartphone development.
In military terms, however, the appeal of smaller doesn’t just mean harder to detect and potentially faster and more agile. It also means designing systems without having to create space for human operators, which saves on protection and life-support systems (armor, windows, and seats on an Army vehicle, for example, or oxygen tanks and an ejection seat on an Air Force plane). These are expensive and heavy.
The Pentagon’s ultimate goal is the creation of “swarms” of small, expendable flying or swimming weapons that can overwhelm adversaries in sheer numbers—which can be programmed to behave precisely like real-life swarms of insects, birds, or fish. A company has already developed a commercial drone in which the joystick operator controls one “smart” leader with four other drones mimicking and following its movements. The next step is creating the software that would enable the leader to direct a fleet of other drones in complex maneuvers that evade anti-aircraft systems or even electronic jamming and that close unerringly on their target. “The power and sheer speed of execution would give them a huge advantage over their adversaries,” according to an article called “The Coming Swarm,” in National Defense magazine.
Still, the key to such future systems rests on another development growing out of the civilian sector: artificial intelligence. This means the creation of machines that can interpret and “learn” from changes in their environment and then communicate a response across a broad network. Such a step would be crucial to developing swarm technologies, or to creating large-capacity “mother ships” that release numbers of other smaller unmanned vehicles while retaining command and control through the network to complete the mission—with or without the need for a human operator.
The same is true of Big Data. Instead of teams of humans staring at computer screens trying to figure out what reams of data or images retrieved from a high-flying UAV or satellite mean, Big Data can generate the algorithms that autonomously separate the wheat from the chaff and allow for faster decisions and better intelligence-gathering. Algorithms can also find unexpected patterns in disparate data that would allow a military commander to pinpoint an opponent’s hidden weaknesses, not to mention his own.
This will be increasingly necessary in dealing with threats like hypersonic attack, or even cyberattack, which will come so bewilderingly fast that effective response by human operators will be impossible. “When you are under attack, especially at machine speeds,” says Work, “you cannot have a human operator, operating at a human speed, fighting back.” Artificial intelligence, reinforced by Big Data, translates into a more responsive and effective defense and offense.
A third, and by no means last, area of commercial technology the Pentagon is actively looking at is 3-D printing (so-called smart manufacturing). The aim here is to move the manufacturing process off the factory floor and directly onto the battlefield. One of the logistical nightmares any army faces is not having the right weapons at the right place at the right time, and having to wait until supplies or reinforcements arrive—which, in a confused battle space covering large distances or in a densely packed urban environment, might be never.
Smart or 3-D manufacturing would actually enable a combat unit to “make” the weapons it needs on the spot. This would include, for example, spare ammunition or mini-drones for reconnaissance—even brand-new weapons more suited to the environment or the fire fight they’re in. In short, “smart” manufacturing processes will change not only the way weapons are made, but also how they’re used and where. The concept of “just in time” manufacturing could become a reality on the battlefield as well as in a defense plant, and render obsolete the need for massive stockpiles of ammunition and other equipment that keep many a supply sergeant or quartermaster general busy—or a little crazy.
Will all these developments, then, spell the end of relying on conventional arms, or conventional supply and logistical chains—or perhaps the end of the need for humans to do any fighting at all?
By no means. The timeline of the Third Offset, even with the most optimistic forecasts, is going to ensure that America will have to rely on its current array of bombs and missiles and its fleets of manned ships, submarines, planes, and M-1 Abrams tanks for years to come. Indeed, part of the plan is to use these emerging technologies to upgrade the arsenal left over from the Second Offset of the 1970s and 1980s, as well as to find new uses for old machines (as when the B-52 shifted from carrying nuclear weapons and conventional bombs to unloading payloads of cruise and precision-guided missiles over Afghanistan and Iraq).
Still, the Pentagon is now poised to reverse a trend of more than three decades of relying on weapons that are big, complex, and expected to perform multiple functions—such as the B-2 bomber, the F-35 Joint Strike Fighter, the Ohio-class submarine, and the supercarrier USS Gerald Ford. All of these come with hefty price tags. The new arsenal will include more and more tools that are small and cheap and will generally perform one function before they crash or burn up or are thrown away. In fact, the Air Force’s term for the drone swarms they envision for future operations is “kamikaze drones”—although in this case there won’t be a pilot who has to die as well.
That’s the chief selling point about the use of “autonomous” systems in battle: No humans will have to die on our side. As systems become more intelligent and more discriminating in selecting targets, the numbers of those killed as part of “collateral damage” will steadily drop, as will the number of enemy combatants who have to be eliminated in order to achieve a victorious result.
The use of “smart” and semi-autonomous weapons like the JDAM “smart bomb” and the Predator drone has already meant a drastic drop in mortality rates in combat situations for both sides in armed conflicts (claims from human-rights groups and others about the “immorality” of Predator drone strikes notwithstanding3). The advent of “intelligent” weapons will reduce those numbers even further. Instead of having to blow up a car full of people with a Hellfire missile to eliminate one al-Qaeda or ISIS terrorist, it will be possible to launch a single miniature drone that uses face-recognition software4 to find its quarry, and which can gently attach itself to his clothing before blowing itself up—or injecting a fatal toxin.
Just as the First Offset strategy of the 1950s and 1960s rested on the “nuclear triad” of strategic bombers, land-based ballistic missiles, and nuclear-armed submarines or “boomers,” so the Third Offset ultimately rests on a “smart triad” of cyber, artificial intelligence, and autonomy—which means ceding more of the decision-making to the machine.
The reliance on machines rather than people, as the main combatants in future wars, makes many uncomfortable, not to say terrified. Yet the fact is, we have already advanced very far down this road, and not only in the military sphere. The car that tells us when our gas tank is low, our door is open, our seat belt isn’t fastened, our tires need rotating, and another car is coming up on our left, is filled with autonomous systems; Google X’s and Toyota’s driverless cars are only the next inevitable steps. Of course failure of any these systems could be dangerous, even catastrophic—for example, a broken gas gauge while crossing the Mojave Desert. Yet we accept that these autonomous systems are there to enhance, not take over, our driving performance, and that by and large they make us safer rather than not (although one might hesitate to be the first passenger in a driverless car on the freeway).
Likewise, virtually every weapons system the military uses today requires some degree of autonomy. That is, they involve the carrying out of certain automatic processes without relying on a human controller or monitor. It is true that the Third Offset will vastly increase the ability of future machines to make more and more decisions on their own, while developments in artificial intelligence will increase the range and sophistication of those decisions. It will even enable machines to “learn” from their environment and seamlessly adapt their range of choices and then pass that data on to other machines in the network, without bothering to inform their human operators.
Will those decisions inevitably include deciding when to shoot to kill? That’s the scenario that scares many who are worried about the autonomy of our “smart” military to come: the killer drone that’s able to find and fix its target on its own, and then strikes without human command—or in the worst-case scenario, even in defiance of that command.
That scenario is unlikely, for various reasons. A 2012 Defense Science Board study entitled “The Role of Autonomy in DoD Systems” dispels the notion that relying on unmanned systems means machines taking over the battlefield. It stresses that the term “autonomous” does not mean “self-governing” or even having the capability for independent thought or action. After all, even the driverless car will have to have some place to go, which means it’s ultimately built to respond to human wishes and commands, rather than vice versa.
“Autonomy is better understood as a capability (or set of capabilities) that enables the larger human-machine system to accomplish a given mission,” the study says. That means letting the machine carry out such tasks as verifying speed and distance, reacting to immediate changes in the environment, self-correcting software or mechanical glitches, and deciding when it would be optimal to advance to the next step in the mission plan. From this perspective, the key advantage to increasing autonomy is that it improves and enhances the “man-machine interface” to the advantage of both.
Both the Defense Science Board study and the architects of the Third Offset assert that even the biggest leaps in artificial intelligence won’t eliminate the need for the human operator and controller, especially on the battlefield. The more potentially lethal the decision point becomes, in fact, the more human controllers will be needed to make the final call—while faster processing of information from the autonomous system will afford them more time to make the right call, just as faster and more cyber connections will mean less opportunity for an opponent to jam or hack an interactive system. Indeed, with the simplest robotic weapons like the mini–sea mine, there would be no cyber connection to hack.
That, at least, is the official position of the U.S. military. Others may not be so choosy. It’s difficult to imagine the Pentagon or a U.S. defense contractor designing an autonomous system that purposely doesn’t discriminate between enemy and innocent bystander. But it’s very easy to imagine a terrorist designing one that does.
This is where the rise of weapons that are smaller, cheaper, and smarter gets frightening. As the National Defense University’s T.X. Hammes concludes in a policy-analysis paper for the Cato Institute, “the convergence of these new and improving technologies is creating a massive increase in capabilities available to smaller and smaller political entities—even the individual,” and it “provides smaller powers with capabilities that used to be the preserve of major powers.”
Cyberwar is one obvious example where a determined hacker can create havoc in networked systems that have become essential to all branches of the military. Drones are another. At minimal cost, a handful of small drones carrying explosive charges can be programmed to hit targets that are proportionally far more valuable and vital (Hammes’s example is insurgents targeting army fuel trucks).
In fact, the real danger is that as Moore’s Law makes high-tech weaponry cheaper, smaller, and smarter, the proliferation of these technologies in the hands of bad actors starts to seem inevitable. A lone-wolf terrorist armed with a drone he can fly and detonate from a remote site is terrifying enough. One who can release a swarm of drones that evade detection and hide before exploding multiplies the possibilities of chaos.
So what’s the answer? Certainly not the response taken by a thousand scientists and researchers who signed a petition, sponsored by the Future of Life Institute in 2015, calling for a permanent ban on autonomous weapons. Like Pope Innocent II’s ban on cross bows in 1139, it is already too little too late—and will be just as efficacious.
Instead, the key might be who gets out of the starting gate the fastest. Certainly the United States can seize what strategists would call its “first-mover advantage,” to take the high ground in developing and perfecting these Third Offset technologies before anyone else does, in ways that will make the economic and political cost of catching and keeping up daunting to competitors. We can also be the first to share the development and deployment of these technologies with allies, which will deliver multiple advantages. These emerging autonomous technologies can enable NATO allies to vastly enhance their military capabilities without having to spend large sums on expensive systems like tanks, helicopters, destroyers, and F-35’s. As for Israel, co-development in areas like AI and cyberspace could be world changers, not only in the military sphere but in commercial spin-offs, as well.5
Those who complain about the high cost of supporting and defending our allies around the globe, including NATO, will need to take a close look at the Third Offset as a way to relieve the burden on ourselves and to make the U.S. and its friends safer.
Of course, taking full advantage of the Third Offset technologies might not stop others from trying to do the same. But in the Brave New World of the intelligent battlefield, who gets there “first with the most” will have the initiative and will set the pace and direction of the arms race to follow—while buying time to start planning the Fourth Offset.