Peace Magazine , page . Some rights reserved.
Search for other articles by kgsimons here
Today is your unlucky day. You hear a terrifying humming sound approaching, and you immediately run away from it as fast as you can. You know that if that sound gets close enough, you’re done for. It’s an AI-piloted kamikaze drone, programmed to hunt and eliminate whatever it identifies as a target. “Why me?” you ask as you run for your life, trying not to get blown up. Maybe you just look too much like a target? Or, maybe it’s not about you. Maybe the drone has
Delegating ultimate authority over human life to an algorithm crosses a red line
simply been set to kill any human in sight, and you were in the wrong place at the wrong time. You see a building you can take shelter in, but it’s too far away and the drone is too close now. There’s no way you can outrun it. Think quick. Maybe there’s a human at least remotely overseeing this drone. As a last resort, you turn around to face the machine and put your hands up in surrender. If there’s someone seeing what the drone sees, they’ll surely make this stop. Too bad there is no human overseeing this. There is only AI. And AI can’t recognize surrender.
Disturbingly, this scenario is not beyond the realm of possibility today. AI-powered weapons are already here. Weapons that use AI to select and attack targets without Human control are in use and have killed. This July, a report revealed that Ukrainian “fully autonomous drones with no human oversight” set to Terminator mode had killed Russian soldiers. Then in August, it was reported that three Ukrainians were recently killed by a Russian AI-piloted drone without human oversight. The concept of an autonomous killer robot is no longer limited to science fiction but is, in fact, part of our current reality.
While AI weapons might sound like a new phenomenon, it’s a topic that has been discussed at the United Nations (UN) for over a decade. AI weapons (or autonomous weapon systems in UN-speak) have been subject to debate and scrutiny
While AI weapons might sound like a new phenomenon, it’s a topic that has been discussed at the United Nations (UN) for over a decade.
at the UN in Geneva, Switzerland, since 2013; and in 2016, the UN officially created a Group of Governmental Experts on Lethal Autonomous Weapon Systems (the GGE LAWS), tasked with examining and addressing the risks of AI weapons.
The reason the GGE LAWS was even created is because experts warned that integrating AI into weapon systems—giving weapons the ability to independently make life-or-death decisions—would pose risks unlike those we’ve ever faced and for which existing international law does not account.
AI weapons are one of the world’s most dangerous applications of AI. Think of general concerns about AI: bias, discrimination, errors, lying, rogue action—and imagine them concentrated in a machine specifically designed to kill. That’s why AI weapons pose significant legal, security, and ethical dilemmas.
Accountability gap. Current international humanitarian law (IHL, or the “laws of war”) only holds humans to account. If an AI weapon commits a war crime, it’s unclear who would be held responsible. The programmer, the operator, no one? AI weapons create a legal accountability gap where war crimes could be committed and merely blamed on AI.
Unreliability and unpredictability. AI cannot reliably identify surrender or injury; and it unpredictably produces errors and fabricates information, often without clear explanation. This not only raises the risk of AI weapons violating IHL, but also makes it extremely difficult to predict whether, or explain why, they violated it in the first place.
Proliferation. Unlike nukes or ballistic missiles, AI weapons are cheaper and easier to access. This means their likelihood of proliferating, and finding themselves in the hands of bad-faith actors, is much higher than that of other highly destructive weapons.
Lower threshold to conflict. Proponents of AI weapons argue that they reduce combatant and civilian casualties due to less need for troops on the ground and higher attack precision; but that narrative simply lowers the political cost of triggering conflict and makes it easier to justify doing so. Rather than reduce harm, using AI in war accelerates the number and speed of attacks, leading to increased death and destruction.
- Moral red lines. AI lacks human qualities—such as morality, empathy, judgement, and accountability—that make people responsible for following laws and norms. AI weapons reduce humans to mere data points in a calculation of whether or not to target and attack. Delegating ultimate authority over human life to an algorithm crosses a red line: AI cannot comprehend the value of human life, so it should never decide who lives or dies.
In light of these dangers, governments at the GGE LAWS have been debating whether to establish new international rules to govern AI weapons and, if so, what they should be. One would think that this is a fairly straightforward task, as opposition to AI weapons and support for new rules are globally popular positions.
So what’s the hold up at the UN?
For years, leading AI experts have sounded the alarm on AI weapons and demanded their ban (see calls from 2013, 2015, 2016, 2017). Similarly, the highest authorities in international governance and humanitarian law—the UN Secretary-General and the President of the International Committee of the Red Cross—jointly called for the urgent establishment of new laws on AI weapons in 2023 and this year. The UN Secretary-General went even further in 2025 and again this year, asking for an outright international ban on AI weapons, calling them “politically unacceptable” and “morally repugnant.”
An overwhelming majority of 130 UN States also officially support the establishment of new international laws on AI weapons. Moreover, polls show time and time again that global public majorities oppose killer robots and want them banned. So what’s the hold-up at the UN?
A minority of governments are still not in favour of (and are delaying progress towards) establishing international legally binding rules on AI weapons—regulations and prohibitions that could ensure limits on AI weapons and the protection of people against them. This group includes countries whose positions remain undeclared and those that categorically oppose new rules. Canada’s position is currently undeclared.
Despite Canada voting in favour of every UN resolution on AI weapons so far (in 2023, 2024, and 2025), which all encourage the advancement of the UN’s work on AI weapons governance, Canada still does not support the establishment of international legally binding rules on AI weapons. This is reflected in its actions at the GGE LAWS: Within the past year and a half, Canada has not endorsed multiple joint statements simply calling for the start of official negotiations on rules (not even specifically legally binding ones!). And just in the last GGE LAWS session, where countries had to finalize a draft of possible regulations and prohibitions for AI weapons, Canada supported weakening or removing parts of the text meant to preserve human safety and human control over life-and-death decisions.
That the concept of unfettered AI weapons deciding who lives and dies is still not a clear enough red line for Canada contrasts starkly with its once-revered moral clarity and global leadership in establishing the international Mine Ban Treaty (the Ottawa Treaty) and the International Criminal Court.
Canada’s “undeclared” position has one silver lining: It means Canadians can still sway Canada’s official position.
However, Canada’s “undeclared” position has one silver lining: it means Canadians can still push Canada’s official position towards support for new international legally binding rules on AI weapons. As a matter of fact, this November, there will be a rare and pivotal opportunity for Canada to finally take a stance.
On the week of November 16–20, 2026, the UN in Geneva will hold a meeting called the Review Conference of the Convention on Certain Conventional Weapons, in which countries will decide whether or not to finally launch formal negotiations on international legally binding rules for AI weapons.
Launching negotiations will require as much support as possible, and every country’s position will count. If the international community misses this opportunity and fails to launch negotiations, it might be years until another opportunity presents itself. That means unchecked killer robots putting populations at risk for longer.
We are dangerously close to living in a world where it could be normal for AI weapons to be unleashed on populations without limits, rules, or accountability.
If you do not want to live in a world like that, urge the Canadian government to support negotiations for new legally binding regulations and prohibitions now. On a question as simple as whether or not we will allow AI to decide who lives and dies, Canadian citizens have the power to draw a red line when the government won’t.
With the upcoming meeting in November, the UN can get one step closer to drawing this red line and delivering a tool to protect humans in a world endangered by AI weapons. Canada needs to take a stand. There is no time to lose.