This story is currently available in англійська only — a translation isn't ready yet.
UN weapons convention opens Geneva talks on killer robots as US and Russia refuse a ban
128 states begin a five-day review conference in Geneva on Monday; Germany and France back a two-tier approach while Washington and Moscow reject negotiations.

The 128 member states of the UN weapons convention open a five-day review conference in Geneva on Monday to try again to bring autonomous weapons under control, with Washington and Moscow refusing to negotiate a ban.
The push follows a strike in late February in which a cruise missile aimed at Iran's Revolutionary Guards hit a girls' school in the city of Minab, killing more than 150 people, most of them schoolchildren. Investigations pointed to time pressure and outdated intelligence, and to excessive trust in AI-assisted targeting planning.
UN Secretary-General António Guterres and Mirjana Spoljaric, president of the International Committee of the Red Cross, warned before the conference that the world stands "dangerously close to crossing a moral red line".
The convention, in force since the end of 1983, bans or restricts conventional weapons that cause excessive suffering. Members have agreed use restrictions on incendiary weapons such as napalm and a total ban on blinding laser weapons, but non-binding talks on killer robots have stalled for around 13 years.
At the review conference, the US under President Donald Trump and Russia under Vladimir Putin opposed opening negotiations, together with Belarus, North Korea and Israel. By contrast, 163 states warned of the humanitarian and security risks of autonomous weapons and demanded rapid international regulation.
Germany and France are pushing a two-tier approach in the UN process, according to the foreign ministry in Berlin: a ban on fully autonomous weapons systems removed from human control, plus regulation of systems with autonomous functions. Lieutenant General Christian Freuding, inspector of the army and designated inspector general, told a BDSV industry event in Berlin on Friday that the type and extent of human control must always fit the situation. A defence ministry AI concept published in August says the degree of human interaction depends on the operational context.
Decisions at the review conference are taken by consensus, giving every state a veto. Mauro Gilli, professor of military strategy at the Hertie School in Berlin, said he is "fundamentally sceptical" that agreement can be reached.
Heiko Borchert, co-director of the Defense AI Observatory at the University of the Bundeswehr Hamburg, said the technologically leading actor wants to keep competitors at a distance, while the laggard wants to stop the leader pulling further ahead. Gilli said the US, Israel and China are the most important developers, while Ukraine and Russia deploy systems with increasingly autonomous functions, and named Anduril and Helsing among new market entrants.
Sources — how we know this
- PrimaryHandelsblatt