The Era of AI-Powered Drone Warfare is Coming
Charles Ferguson, the editor responsible for artificial intelligence at Project Syndicate, is a technology investor, policy analyst, and director of several documentaries, including the Oscar-winning film Inside Job.
The Ukrainian attack on air bases across Russia on June 1 signaled the emergence of a new prevailing opinion: the costly human-operated weapons (tanks, planes, ships) that have long distinguished the "advanced" armies of the world are now obsolete due to inexpensive drones. However, this view is incomplete and potentially dangerously misleading. The current drone warfare offers real lessons that go far beyond the obsolescence of outdated, expensive weapons; the imminent integration of artificial intelligence into drone warfare will make the present situation seem positively quaint.
Let’s reflect on the lessons learned from the Ukrainian war so far. First, the impact of drones far exceeds that of older weaponry. Drones have rendered tanks and armored personnel carriers extremely vulnerable, which is why Russian ground attacks often rely on foot soldiers, motorcycles, or all-terrain vehicles. But this has not been effective, as drones are terrifyingly effective against personnel as well. Human casualties remain as high as ever—drones are now responsible for over 70% of casualties on both sides.
Drones are also effective against almost everything else. Ukraine has deployed drones to destroy a variety of Russian targets, such as arms factories, moving trains, ammunition depots, oil refineries, ships, and ports. In fact, Ukraine has demonstrated considerable restraint given Russia’s brutal actions. Airports, train stations during peak hours, sports stadiums, concert venues, pharmaceutical factories, hospitals, schools, and nursing homes are all equally at risk.
Two additional realistic lessons from Ukraine pertain to how drone warfare depends on its industrial base. First, speed and responsiveness are crucial. The technology of drones, their weapons, and tactics are now advancing at a breathtaking pace. A new drone will be useful for only two to six months, as the opposing side develops countermeasures, necessitating the development of new products, which, in turn, face new countermeasures, and so forth.
Initially, the drones used in Ukraine were primitive weapons, controlled via radio by a pilot who needed to be nearby. As drones became more sophisticated and lethal, jamming techniques were employed to block their radio signals. This resulted in frequency shifting, then hopping between frequencies, leading to the use of multi-frequency jammers, which prompted the emergence of drones that would attack jamming equipment. Russia then developed drones controlled via fiber optic cables—resistant to jamming. Now, Ukraine tries to trace the cables back to their source to eliminate the operators (using drones), and Ukraine also has fiber optic drones now.
Guidance technologies are also undergoing continuous development, allowing drones to evade radar by flying very low or using stealth technology. However, drone detection and tracking systems have also advanced, employing networks of mobile phones and microphones connected to sound analysis software, along with technologies like Lidar, radar, and cameras.
In this intense environment, even a one-month delay in keeping up can be fatal. Traditional defense industry measures have become completely inadequate, and most drones and their manufacturers in the United States have proved to be slow, costly, and almost useless. In response, the drone industry and the Ukrainian military have developed a revolutionary model for weapons research, development, and deployment, based on direct and continuous communication between frontline units and drone manufacturers. The Ukrainian military leadership and the Ministry of Digital Transformation even developed a point-based system that continuously updates rankings of military units’ performance based on kills achieved by drones.
Here, Ukraine benefited from a strong ecosystem of startups supporting the arms industry (with hundreds of companies) capable of designing, producing, and delivering new weaponry within weeks. This year, Ukraine will produce more than four million drones, most of them models that did not exist even a year ago. Unfortunately, Russia has also adapted, relying heavily on private startups.
The drone warfare in Ukraine presents another lesson for the United States and Europe: the need to counter Chinese dominance over the global drone industry.
Ukraine developed its own drone industry because the United States and NATO had virtually no industry of their own, let alone one with the necessary speed and agility, and due to China gradually reducing supplies to Ukraine in favor of Russia. In fact, about 80% of the electronics used in Russian drones come from China. While Ukraine initially relied heavily on China, it has reduced its dependence to perhaps 20%, most of its needs now sourced covertly.
Nevertheless, U.S. and European defense R&D and procurement processes remain slow and unable to compete, crippling their ability to defend against drones, as well as to use them. While few people realize it, the United States and NATO urgently need Ukraine’s expertise in drones. Ukraine is now the only country that can match Chinese and/or Russian technology and response times in warfare. Without Ukraine, and without modernizing their own forces, NATO and the United States could suffer devastating losses in any war with Russia or China—and might even lose.
Moreover, artificial intelligence will change everything. In Ukraine’s June 1 operation, 117 drones were deployed, each controlled by a skilled operator. Reports suggest that nearly half were defeated by Russian defenses—primarily jamming—because the drones needed to be wirelessly connected to their operators. If these drones were autonomous, there would have been thousands of them. With artificial intelligence, the need for communication with pilots disappears, thus rendering jamming ineffective, which significantly enhances the range and lethality of drones. Five years from now, launching preemptive strikes on traditional targets will be frighteningly easy.
Artificial intelligence also enhances the lethality and accuracy of drones used against individuals. Researchers have already demonstrated groups of drones navigating through forests and then regrouping as a swarm after passing through. This isn’t just about warfare; it also applies to terrorist attacks.
It is true that the AI functions required still need far more computing power and memory than can fit in a small drone. They are also not cheap. For instance, Nvidia chips can cost up to $50,000 each, making a powerful AI processor prohibitively expensive for most drones.
But this is rapidly changing, driven by the goal of embedding serious AI capabilities in every smartphone. When this happens, those same capabilities will be available in every type of drone. With the sole and vital exception of AI processors, China dominates the entire supply chains for both smartphones and drone weapons.
Stuart Russell, an AI expert at the University of California, Berkeley, has long advocated for a treaty to prevent the proliferation of small, mass-produced autonomous drones controlled by AI. He even produced a short film titled Slaughterbots, depicting the risks these drones pose if they fall into the wrong hands. He told me at a dinner years ago that targeting individuals using facial recognition technology would soon be quite easy, or targeting anyone wearing a cross, yarmulke, or any other religious or political symbol.
Since reaching any real treaty is unlikely in the current geopolitical climate, we must prepare for a world that may contain such weapons. But the Western defense establishment increasingly appears as a typical "antiquated" company caught off guard by a new, disorienting technological evolution. In markets, resistance based on legacy can be costly, but the costs are purely financial. In warfare, they can be deadly—and they will be.
