US and Russia move to block binding rules on killer robots at UN talks

A diplomatic retreat kicks off Friday in Geneva, setting the stage for the seventh review conference on certain conventional weapons, running 16 to 20 November with 128 countries in the room. The question on the table: does the world finally commit to negotiating hard rules on autonomous weapons, or settle for another round of voluntary chatter? Russia, the US and Israel are among the states working to make sure it's the latter, according to The Guardian. They want no timetable, no mandate, no binding text.

The stakes are not abstract. Negotiators have spent nearly three years inside the UN on this, and Anna Hehir, head of military AI governance at the Future of Life Institute, says states have already agreed by consensus on elements of a draft treaty text that would anchor talks. Her warning: expect some delegations to hide behind language that looks like progress while quietly pushing years more of non-binding discussion. Those proposals, she said, are smokescreens to keep autonomous weapons unregulated and states unaccountable.

The pro-regulation camp is loud and growing. On 22 October a coalition of 50 states — Ireland, Italy, Brazil, Mexico, France and Germany among them — called for the conference to agree a strict negotiating mandate. The International Committee of the Red Cross said separately that more than 70 states back opening negotiations, and 163 states voted at the UN disarmament and international security committee that they're ready to launch talks. Ireland's proposal would create a time-limited expert group negotiating rules over three sessions in Geneva in 2027, totalling 25 days, with a draft protocol delivered to that year's annual meeting. The UK didn't sign on to the Irish statement and has argued existing international humanitarian law is enough.

High-level voices aren't staying quiet. In an intervention timed to sway Geneva, Pope Leo XIV said decisions affecting human life, freedom and the use of force must stay under effective human oversight, and urged states to take concrete steps toward banning lethal autonomous weapons. In August, outgoing UN secretary general António Guterres and ICRC president Mirjana Spoljaric warned the world was nearing a moral red line where autonomous weapons could kill without human involvement. On Thursday the ICRC put it bluntly: the time has come to move from discussion to negotiation, and the world cannot afford further delays.

Battlefield reality is moving faster than the diplomacy. In July, Ukraine reported for the first time that Russia used an AI-guided drone, striking a propane tank at a fuel station in Zaporizhzhia; Volodymyr Zelenskyy has warned about the pace of AI weapon development and called for the war to end before AI takes over the battlefield. The Guardian also cites the 28 February US bombing of a school in Minab, Iran, which killed at least 175 people. The targeting involved input from Palantir's Maven AI system, which reportedly didn't flag that the site had been a school for a decade — and no reviewing personnel caught it either.

For the record, the expert working group defines autonomous weapons as those capable of identifying, selecting and engaging a target without referring the decision to a human. A Future of Life Institute-commissioned poll found overwhelming public support for a global treaty in France, Germany and the UK, though plenty doubt it would ever be respected. With 163 states already signalling they'd negotiate, the pressure to fold and accept a deadlocked outcome will be enormous. History says arms control deals happen when cross-regional majorities refuse to blink.