When rogue fighters abandoned their months-long siege of Tripoli in the spring of 2020 and scattered into the Libyan desert, something pursued them from above, hunting them across the open terrain: drones. Built by Türkiye, these machines were designed to keep flying even if they lost contact with their human handlers. In other words, they could identify and attack targets on their own.
We put ‘u’ in neighbour and you in Canada’s conversation. Enjoy a roundup of Canada’s best writing. Sign up for The Walrus newsletter and get trusted Canadian journalism straight in your inbox.
The episode came to light a year later in a 548-page report submitted by a United Nations panel of experts to the Security Council. The panel had been investigating Libya’s descent into chaos after the 2011 overthrow of long-time dictator Muammar Gaddafi created a power vacuum that unleashed years of fighting between rival governments, militias, and warlords. UN reports rarely generate global headlines. This one did because it appeared to offer the first documented case of a supposedly lethal autonomous weapons system—the kind of killer robot that until then was confined to science fiction. Whether this function was ever used on the Libyan fighters remains a mystery, but it portended how armed combat was about to evolve.
Kyiv first fielded fully autonomous drone capabilities—“Terminator mode”—against Russian soldiers in 2024. The conflict’s front lines have since become a laboratory for relentless automation. Quadcopter first-person view drones—cherished by wedding planners and travel influencers—have morphed into marauding executioners hurtling themselves toward enemies at 150 kilometres an hour.
Ukraine is retaking chunks of its territory from Russia by launching all-robot assaults, posting videos of Russian infantry staring into the camera as they surrender to drones and unmanned ground vehicles. The Kremlin has established its own military branch dedicated to autonomous systems as well. Russia’s hulking war machine may be battered, but it also now has more expertise in high-tech combat than any of its autocratic peers or Western adversaries.
Elsewhere, Israel has deployed intelligent military systems in Gaza and plans to do the same along its borders with Lebanon and Syria. Chinese killer robots were openly paraded through the heart of Beijing last September. The Donald Trump administration in the United States is demanding Congress greenlight $54.6 billion (US) in 2027 for the Pentagon’s new Defense Autonomous Warfare Group. Autonomous defence tech firm Anduril (a darling of Silicon Valley venture capitalists) wants Washington to ease export restrictions and allow it to set up branch plants abroad.
It’s a race for arms now small enough to fit in the palm of a soldier’s hand. In 2023, an Indian defence start-up unveiled three variants of nano drones, each weighing about 200 grams, or roughly the same as a large apple. Once programmed, a kamikaze model is capable of identifying an enemy’s hideout and self-detonating upon arrival.
“It used to be that only the strongest countries could use drones—because they were expensive and complex,” President Volodymyr Zelenskyy told the UN General Assembly last fall. He warned a future of fully autonomous machine warfare is inevitable. “Now, even simple drones can fly thousands of kilometres. War tech doesn’t care about geography anymore—it’s now reshaping it.”
The consequences could be phenomenally grim. The international community is grappling with how to mitigate the risks. Efforts to craft even imperfect safeguards—similar to those placed on nuclear warheads, chemical agents, anti-personnel mines, and cluster munitions—are essentially deadlocked. Civil society groups and human rights organizations are pushing for idealistic prohibition treaties. Military heavyweights want a maximalist approach.
Most countries sit in the middle, hedging their bets in a geopolitical environment that is reverting to the law of the jungle. Canada is no exception. Ottawa’s Defence Industrial Strategy is funnelling billions into the sector, pledging to expand the domestic capacity of unmanned systems and establish a Drone Innovation Hub. Several major barriers stand in the way. But if they can be overcome, Canada has a rare opportunity to show what the responsible use of the technology looks like in practice. For Branka Marijan, a military researcher specializing in emerging technologies, that means creating “safeguards other countries can actually adopt.” It’s the kind of leadership in desperately short supply.
Not all autonomous weapons systems are the same. Some, such as self-guided ballistic missiles or Israel’s Iron Dome air defence network, have been around so long they’ve been normalized. But recent leaps in artificial intelligence, machine learning, robotics, and visual recognition have ushered in a vast spectrum of new options. For the most part, they are used to conduct intelligence, surveillance, and reconnaissance (ISR) missions. And humans generally remain “in the loop” on life and death decisions.
That is changing fast. A 2025 report from the UN secretary general forecasts the current rearmament bonanza may reach $6.6 trillion (US) by 2035. That’s more than doubling the record-shattering outlays of $2.7 trillion (US) in 2024. A growing portion of these titanic sums has already manifested a fearsome arsenal of autonomous weapons systems that seem destined to be granted ever more agency—aerial drones and ground vehicles, armed sentry outposts and gun-toting mechanical dogs, uncrewed marine vessels and AI “wingmen” that amplify the firepower and reach of warplanes. Foundation, a San Francisco–based company with hefty Pentagon contracts, is trialling humanoid robot soldiers in Ukraine. China has unveiled compact unmanned attack submarines.
All of this is happening because the strategic benefits are undeniable. “Both the war in Ukraine and the war over Iran are shaped by technology which has introduced a new transparency to the places and situations in which armies fight,” wrote Shashank Joshi, The Economist’s defence editor, in May. To him, the ability for drones, satellites, and battlefield sensors to generate real-time intelligence and targeting data has become the defining trend of modern war. Any person or piece of equipment that exhibits even the faintest movement can be easily tracked and swiftly eliminated with devastating precision.
Waves of cheap robots also invert the economics of war by enabling a smaller actor to burn through a larger enemy’s stockpiles of expensive munitions and legacy platforms. A $1,000 drone can destroy a tank costing millions. Advances in design, connectivity, and propulsion mean enemy targets and resupply positions deep behind the front lines are increasingly vulnerable too. Possessing legions of expendable weapons systems eases military personnel pressures as well.
Former US military commander and Central Intelligence Agency director David Petraeus has predicted that swarms of killer robots could coordinate into battalion units without direct human intervention. In future conflicts, he’s written, “the side that waits for human approval before acting will lose.” And he’s not alone in thinking this way. The British military is reportedly reconsidering its long-standing prohibition on autonomous lethal strikes, arguing that defeating fast-moving adversaries in extreme battlefield conditions may require removing people from the decision-making loop altogether.
