Killer Robots Have No Rules — and This Week in Geneva Decides Whether They Ever Will

There is a class of weapon that can already find a target, decide it is a threat, and kill it — without any human making the final decision. There is no treaty that governs it. This week, for the first time, the world gets to decide whether to start writing one. And the two countries that build most of these weapons are expected to say no.

Autonomous military drone and ground robot silhouetted against a pale sky — Global Insight

According to reports from Reuters and the Intelligence Report, the Seventh Review Conference of the Convention on Certain Conventional Weapons opened on Monday, November 16, at the Palais des Nations in Geneva, under the presidency of Ambassador Julien Thöni of Switzerland. It runs until November 20, and it has one decision to make: whether to launch negotiations on a legally binding agreement covering lethal autonomous weapons systems — LAWS, the diplomat's term for what everyone else calls killer robots.

What exactly is a lethal autonomous weapon?

The definition matters, because the whole negotiation turns on it. A lethal autonomous weapons system is one that can select and attack targets without a human making the final decision. That does not mean a science-fiction robot soldier. It means the drones, robots, tanks and submarines that artificial intelligence has already made capable of killing without human involvement — systems that identify a target, weigh it against programmed criteria, and fire. The line the negotiators draw is between a weapon a human aims and a weapon that aims itself.

The Convention on Certain Conventional Weapons, the framework under which the talks sit, is a 1980 UN-hosted treaty that restricts weapons considered excessively injurious or indiscriminate. It has 128 High Contracting Parties and five protocols — covering everything from landmines to blinding lasers. None of its protocols addresses autonomy. The conference will consider the final report of the Group of Governmental Experts on Lethal Autonomous Weapons Systems, which set out "elements for an instrument" — the building blocks of a treaty, without yet being one.

Who wants rules, and who is blocking them

Most of the 128 states party to the convention support launching negotiations on a binding agreement. The campaign group Stop Killer Robots says 130 countries now back a legally binding instrument, though reported counts range from about 76 to 130 depending on the forum measured. The opposition is smaller but includes the countries that matter most militarily: the United States, Russia and Israel. Their argument is that existing international humanitarian law is already sufficient to govern these weapons, and that new binding rules are unnecessary.

The split is not just about whether to negotiate, but about what the negotiation would produce. The conference chair has proposed a two-year mandate to negotiate a legally binding protocol. The United States has proposed the opposite: five years of talks aimed at a non-binding agreement, favouring guidelines over binding rules. The United Kingdom, Germany and France sit between the two, proposing negotiations on an instrument that leaves open whether it would ultimately be legally binding. The UK government declined to back a binding treaty in the days before the conference opened. Russia has not publicly stated its position going into the talks, but it has previously resisted binding proposals — and, according to a senior Western diplomat quoted by Reuters, it plans to block the European Union, an observer to the talks, from taking the floor.

That matters because of how the convention works: it operates by consensus, so opposition from even a few states can block a negotiating mandate. That is why treaty backers are already discussing a route through the UN General Assembly, where a majority vote — rather than consensus — could launch negotiations. The Secretary-General of the UN urged states, in a video message as the talks opened, to agree to move toward a legally binding instrument with clear prohibitions and restrictions. The International Committee of the Red Cross and Pope Leo XIV have also warned of the dangers and support regulation. The Red Cross's point is the one the opponents have to answer: humanitarian law was written for weapons humans control, and it does not fully address the specific risks of systems that make life-and-death decisions at machine speed.

Why this week is the fork in the road

The stakes are not abstract. The same AI advances that produced deepfake and crypto-theft campaigns are now embedded in targeting systems, and the line between a drone a human pilots and a drone that hunts on its own is already blurring on real battlefields. Governments are responding to that pressure from every direction — Germany has just given its intelligence services new powers to hack and disrupt — but autonomous weapons are the one area where no rules exist at all, and where the window to write them is closing as the technology spreads.

By Friday, November 20, Geneva will have produced one of three outcomes: a mandate to negotiate a binding protocol, a mandate for non-binding guidelines, or nothing — another round of discussion, with the decision deferred. If the major military powers block the mandate, the campaigners' fallback is the General Assembly, where the numbers favour a treaty but the biggest weapons-builders can simply refuse to join it. Either way, this is the week the question stops being theoretical. For more than a decade, diplomats have debated whether machines should be allowed to decide who lives and who dies. This week, they have to decide whether they are even going to try to stop it — and the answer will shape warfare for the rest of the century. The honest advice, as with any threat that arrives faster than the rules, is the same one that applies to every other crisis: prepare before you need to, because after is too late.

Post a Comment

Previous Post Next Post