In a new piece in The Atlantic, former Pentagon officials Raj Shah and Christopher Kirchhoff join the chorus of techno-enthusiasts who believe that pilotless weapons and systems incorporating artificial intelligence are the future of warfare.
Shah and Kirchhoff, who ran the Pentagon’s Defense Innovation Unit before going into the investment and consulting worlds, are co-authors of a new book, “Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War.” The Atlantic piece is written in that spirit. It touts the value of new technologies for Ukraine’s efforts to stave off the Russian invasion of their country, and it even goes so far as to say that “America’s technological genius is one of its best tools it has for keeping the peace.”
The notion of “peace through technology” has a long and dubious history. From the “electronic battlefield” in Vietnam to Ronald Reagan’s quest for an impenetrable missile shield to the emergence of precision-guided munitions over the past two decades, new technologies that were supposed to “revolutionize” warfare either failed outright or were of limited utility in the kinds of wars they were being used to wage. Whether the new wave of technology coming out of Silicon Valley meets the same fate remains to be seen.
There are already reasons to doubt whether a future marked by renewed U.S. dominance fueled by the superior development and employment of pilotless and artificial intelligence-driven weapons will ever come to pass. A recent Wall Street Journal investigation of the use of small U.S.-supplied drones in Ukraine found that the systems failed miserably. They were too brittle and too expensive, so much so that they were abandoned in favor of cheaper, more reliable Chinese drones.
The dismal performance of the small drones in Ukraine has undermined the argument of companies and investors involved in developing next generation military technologies that their systems can play a decisive role in combat. Some of the new weapons work, and some do not. The idea that they have all been “proven in combat” is premature, at best.
A corollary of the certainty of the new age militarists that weapons like pilotless vehicles and systems controlled by AI will give the United States a decisive military advantage is the notion that they will be key to “beating China” in a potential conflict or deterring Beijing from greenlighting aggressive action against their neighbors. Deputy Secretary of Defense Kathleen Hicks is a firm believer in this idea, as evidenced in a speech she gave in August of 2023 to the National Defense Industrial Association, the country’s largest arms industry trade group.
Advocates of pouring funds into Silicon Valley to speed the development of fast, nimble, cheap, and easily replaceable new weapons systems rarely speak of the risks involved in going down this path. Any defense system built with incredibly complex software is likely to fail at some point. If it occurs in an AI-driven system, it could cause unintentional mass slaughter. And if AI is ever embedded in the command and control system for nuclear weapons – an approach the Pentagon currently says is not in the cards – the possibility of a system failure that triggers an accidental nuclear weapons launch cannot be ruled out.
There are three possible outcomes for the kind of high tech future endorsed by Shah and Kirchhoff. First, the weapons don’t work as advertised, and we waste tens of billions of dollars while falling behind in modernizing the U.S. military. Second, the weapons work as advertised, and enhance U.S. performance on the battlefield while forcing potential adversaries to think twice about attacking the U.S. and its adversaries. Or third, the weapons work but are employed in ways that cause unintended killing and provoke unnecessary wars. No one knows for sure which of these outcomes will come into being, but option two would run counter to the experience of the vast majority of conflicts of the past 100 years.
At a minimum, we need a vigorous national debate about whether and how to use emerging technology for military purposes, including how to test and monitor new systems for effectiveness and affordability, and what strategy they will be used to implement. This will be a challenge in the face of the kind of techno-enthusiasm displayed by military leaders and segments of the broader public alike, and the fact that venture capital firms and arms company startups are rushing to cash in on the development of AI-driven weapons and other new systems. There is a real danger that financial interests will overwhelm the public interest in questions regarding the development and use of emerging technologies for military purposes. If that happens, it will deal a blow not only to our democracy but to our safety and security, likely for decades to come.
