Militaries around the world are increasingly relying on artificial intelligence to find, track and annihilate targets. An upcoming conference in Geneva offers a chance to confront the gravest risks of AI-enabled warfare, ideally through a treaty to regulate autonomous weapons systems.
A strike that might have taken half an hour and required significant human consideration only a few years ago can now take minutes, or even seconds, with AI systems taking the lead if they meet a confidence threshold set by a human operator.
In Iran, the US relied on AI-enabled targeting to strike 1,000 targets in the first 24 hours of its attacks in February, during which a school in Minab was bombed, killing more than 120 children. The Russia-Ukraine war has seen extensive use of AI-enabled drones with little or no human input in the final stages of a strike.
Across the Middle East, the Israel Defense Forces have used their “Lavender”, “The Gospel” and “Where’s Daddy?” systems to identify alleged militants and their homes as potential bombing targets. Lavender assisted in the targeting of 37,000 Palestinians in Gaza in the months following 7 October 2023, bypassing what a senior Israeli commander chillingly described as a “human bottleneck for both locating the new targets and decision-making to approve the targets”.
These instances are not now outliers, but rather indicate a wider global trend. This now requires a commensurate global policy and legal response to ensure that human beings and states are held accountable for military decisions – not machines or AI systems.
It can seem daunting, if not impossible, to challenge superpowers’ military doctrines and impose tough regulation on weapons systems. But I know from personal experience that it can be done.
Nearly 40 years ago, I played a role in the passage of the legislation which declared New Zealand to be a nuclear-free state. The New Zealand Nuclear Free Zone, Disarmament, and Arms Control Act 1987 to this day bans the entry of all nuclear weapons and nuclear-powered vessels to New Zealand, and has enjoyed a consistently high level of public support since it was first introduced.
Today, we need a comparable level of global ambition in the face of morally and legally complex new challenges.
When an autonomous weapons system assists in committing a war crime, accountability becomes difficult to establish. The risks are particularly acute for anti-personnel weapons – which are designed to target and strike people, rather than vehicles or buildings. These weapons struggle to distinguish between combatant, journalist, civilian or child.
As Pope Leo XIV stated in his recent encyclical, Magnifica Humanitas: “Any technology that facilitates attacks without seeing the face of human beings lowers the moral threshold of conflict”.
Despite these clear and growing concerns, those governments with the most advanced autonomous weapons systems have long resisted binding limits on their development and use. While they have dragged their feet, the technology, data-processing systems and related surveillance tools have continued to advance at breakneck speed. The gap between technological capability and regulatory constraint is widening day by day.
Given that some of these technologies are cheap to produce and easy to copy, there is a significant risk that non-state actors, including terrorist groups, could obtain them. It is in all states’ interests to avoid such a catastrophic outcome.
after newsletter promotion
When states meet this week at the seventh review conference of the Convention on Certain Conventional Weapons in Geneva, they face a choice between launching negotiations on treaty provisions covering these weapons, or allowing more autonomous weapons on to our battlefields and beyond.
Simply agreeing to negotiate a treaty would be a positive outcome given the fractious state of international relations today. With the right level of political will and ambition, these talks can lead to genuinely transformative outcomes, as we have seen in previous decades with binding agreements to ban the use of chemical weapons and anti-personnel landmines.
A meaningful treaty would prioritise human control, accountability and non-proliferation. It would ban autonomous weapons systems which cannot comply with international humanitarian law. All other autonomous weapons would be regulated, including through limited engagement and targeting, human supervision and testing and evaluation.
Those countries which benefit most from the existing lack of accountability may seek to thwart efforts to negotiate a meaningful treaty. That makes it all the more important that the majority of states act collectively to create a safer future, through perseverance, cooperation and courage. A weak treaty which placates those resisting accountability cannot do that.
Technology will not stand still while diplomats struggle to reach consensus; if states do not seize the moment now, our future collective security risks being ever more endangered.


