Twelve years ago, Frank Kendall, then under secretary of defense for acquisition, technology, and logistics at the Department of Defense, pulled up a PowerPoint slide to show a group of senior military officials who deliberate over budgets. It showed the choices he thought the military had to make in space, since China and Russia were working on weapons to take down American satellites that provide everything from intelligence to communications to navigation.
The country, Kendall said, could either keep building the relatively few hyper-expensive satellites of yore, easily attacked and hard to defend; dream up a more resilient satellite infrastructure; or lean back from dependence on satellites. “Business as usual was not a viable choice,” Kendall writes in “Lethal Autonomy: The Future of Warfare Whether We Like It or Not,” an account of a military future where vehicles and weapons operate largely on their own — including making some decisions to strike people or infrastructure. Spacecraft, living in orbit as they do, already work more independently than other systems.
BOOK REVIEW — “Lethal Autonomy: The Future of Warfare Whether We Like It or Not,” by Frank Kendall (Knox Press, 288 pages).
Ultimately, a study recommended reimagining military space, primarily through creating smaller, cheaper satellites that collaborate and can continue their job if a few are taken out, or can be relaunched quickly. The military is still working on that plan, and in Kendall’s view, it’s necessary: Conflict is coming for space.
Conflict, though, is also changing on the ground, and should change even more, Kendall’s book argues. “We are living in an age of exponential technology development that is certain to have a profound effect on how any future conflicts will be fought,” he writes. “We are also living in an age in which a brief period of American dominance in conventional warfare may be coming to an end.”
The U.S. must change to maintain military superiority, Kendall argues, and in the book he lays out a way to reconstitute a military focused on what the book’s title implies: vehicles and weapons that can operate mostly on their own, able to take lives and destroy other robots and infrastructure, while also saving human lives by subtracting soldiers from the battlefield.
“I continue to believe that the appropriate response is to master technological change rather than resist it,” writes Kendall, who was secretary of the Air Force from 2021 to 2025. But the issues of whether that mastery is something U.S. residents and the military want, whether it’s ethical, and how such a shift would change the nature of war, aren’t settled yet.
Kendall envisions a future where vehicles and weapons operate largely on their own — including making some decisions to strike people or infrastructure.
As innovations like self-navigating technology and automated image analysis have increased, the ability for weapons to kill or destroy under their own power has become a nonfictional reality. According to Kendall, the U.S. military has been reluctant to embrace and move forward with those possibilities.
There are many arguments against lethally autonomous weapons: flaws in targeting, the potential for rogue behavior, and a lack of accountability for such mistakes. But humans also make those mistakes. “Is there some good reason,” Kendall argues, “why we’d prefer that humans kill and be killed in war as opposed to machines assuming these functions?”
After all, he continues, history is full of unpleasant tasks being relegated to machines. “It may sound fanciful, but if we have to resolve our conflicts through competitive violence, why not let our machines do that for us?” he writes. “This might strike some as ridiculous, but what is really ridiculous to me is that we still have a preference for sending humans out to kill each other to resolve our differences.”
If weapons systems are as reliable or more reliable than humans in war, and if human officials are held accountable when machines make mistakes, there is, Kendall argues, an ethical case for a largely robotic military, aided by humans.
In “Lethal Autonomy,” Kendall lays out what such a bot military would look like. The book is not pleasure reading — it parses more like a how-to manual. It’s structured according to how different domains of the military would look, were it not centered around human combatants, then laying out the complications in each case. Its pages seem meant primarily for military strategists or those interested in the details of that strategy, and also, likely, as a warning for those in other countries, who can also buy books and see how retired American leaders envision the future of the world’s most powerful military.
“I continue to believe that the appropriate response is to master technological change rather than resist it.”
On land, for instance, Kendall envisions crewless aerial systems that can take off and land vertically. These drones, working together in small units, gather data for automated decision-making, detecting and identifying things the data suggests they should destroy, and then doing so. They’re launched from a driverless ground vehicle that itself contains units of such aircraft.
The ability to identify appropriate to-be-destroyed objects and people, while avoiding collateral damage, Kendall writes, should be testably at least as good and probably better than human-driven intelligence. And it wouldn’t be suitable for all situations — a no-go for, say, a place where children are throwing grenades. In that case, a remote human could get involved; in most of Kendall’s plans, in fact, humans retain some decision-making power.
On the sea surface, Kendall asks for the military to rethink what big ships are even for, since airplanes, satellites, and oceanic monitoring stations can do surveillance, and because land-based weapons can now reach enemy navies.
Under the ocean, Kendall believes crewed submarines are not long for this world, as new detectors make it harder for them to stay stealthy. “Time and technology do not seem to be on the side of the submarines,” he writes.
Instead, Kendall suggests groups of medium-sized autonomous, stealthy subs that can engage an enemy’s submarine together, and — unpeopled — take more damage without loss of life while doing damage.
Get Our Newsletter
Kendall’s true expertise, though, flies through the air. When he was secretary of the Air Force, he initiated the Collaborative Combat Aircraft program, which is developing drones that autonomously fly next to piloted fighter jets and can take direction from the nearby pilot. It’s a start, but Kendall depicts a future with even fewer pilots — focused partly on “unmanned stealthy tactical aircraft” that can take off without long runways and work together.
The difficulty there, he notes, is that automated flight systems are mostly used and will continue to be developed in a commercial or peace-like context. “Combat-related automated behaviors will be much harder to verify,” he admits.
In the air, in and on the sea, and on land, Kendall’s ideas for a robot military are obviously different from the status quo. But they also are basically new versions of conventional warfare. Where “Lethal Autonomy” gets truly interesting is when it veers into orbit, a place Kendall also oversaw in the 2020s since the Space Force is overseen by the Department of the Air Force.
Space, as noted, is pretty autonomous already. But so far, there hasn’t been much lethality in orbit. And there, lethality means killing capabilities that are key to life on Earth, harming satellites that provide essential data, communications, and navigation.
The country, in Kendall’s view, should be space-lethal to adversaries’ satellites when necessary, and able to protect its own spacecraft from attacks. For the former, he recommends anti-satellite weapons; for the latter, perhaps a wartime reserve of satellites, ready to launch at a moment’s notice if something bad happens to their cousins.
The book’s final, and even less conventional, domain is cyber, where Kendall explores the idea of a new, focused, and independent organization. “Unlike everything else I’ve discussed so far for other domains, where autonomy removes emotion from the equation, this form of future warfare is all about psychological impact on the enemy’s population and leadership,” he writes.
The effectiveness of different kinds of attacks would be different in different countries. In China, for instance, the government is concerned about control of information. Decisions about what’s most effective could be helped by an “AI-aided psychological analysis of the targeted society and leadership.”
That, for sure, sounds dystopian. But so does a group of men sitting in a stuffy conference room deciding on operations themselves, as has happened for much of military history.
“I hate war,” Kendall writes. “It is the most wasteful and cruel of human endeavors.”
Flying, floating, submerging, orbiting, rolling, killing machines sound dystopian too. But Kendall’s implied counterpoint has a point: Does a world full of similar vehicles operated by young people, making their own lethal decisions or carrying out those made above them, sound less dystopian — or just more familiar?
In some ways, the concept of an autonomously lethal military demonstrates the ridiculousness of war: Machines owned by one country assert dominance over another country by killing its machines and people, demonstrating a power that perhaps has nothing to do with the substance of the actual conflict.
And, in that, Kendall would agree with you. “I hate war,” he writes. “It is the most wasteful and cruel of human endeavors.”
In fact, he ends the book with the hope that his blueprint is not needed “and the future of warfare is a null set.”
Where might machine-on-machine conflict take us? Kendall asks near the end of the book. “Perhaps the folly and irrationality of using violence to resolve our differences will become clearer and more compelling,” he writes. “I fear, however, that the result will be the opposite.”