Russia stalls UN bid to regulate AI 'killer robots' as talks near deadline
The effort was a possible negotiation on rules for lethal autonomous weapons, discussed at the United Nations’ Review Conference of the Convention on Conventional Weapons in Geneva. These weapons can select and attack targets without human intervention, raising concerns about civilian safety and compliance with the rules of war. Russia slowed the talks in several ways. On Monday, it objected to observers, including the European Union and civil society groups, taking part in the debate. Discussions then continued informally. On Thursday, Russia also demanded repeated changes to the draft final report, including passages unrelated to the core issues. The 128 countries were trying to decide whether to open negotiations and, if so, what terms and timeline to set. Any decision required full consensus. Most countries favoured a legally binding agreement, but states with the technology, including Russia and the United States, were reluctant. By Friday afternoon, the deadlock remained, though one diplomat said countries were close to agreeing that negotiations were needed.
What UN effort did Russia help stall, and what decision were countries trying to reach?
The effort was a possible negotiation on rules for lethal autonomous weapons, discussed at the United Nations’ Review Conference of the Convention on Conventional Weapons in Geneva. These weapons can select and attack targets without human intervention, raising concerns about civilian safety and compliance with the rules of war.
Russia slowed the talks in several ways. On Monday, it objected to observers, including the European Union and civil society groups, taking part in the debate. Discussions then continued informally. On Thursday, Russia also demanded repeated changes to the draft final report, including passages unrelated to the core issues.
The 128 countries were trying to decide whether to open negotiations and, if so, what terms and timeline to set. Any decision required full consensus. Most countries favoured a legally binding agreement, but states with the technology, including Russia and the United States, were reluctant. By Friday afternoon, the deadlock remained, though one diplomat said countries were close to agreeing that negotiations were needed.
What are lethal autonomous weapons systems, and how can they select and attack targets without a person intervening?
Lethal autonomous weapons systems, or LAWS, are weapons that can select and attack targets without a person intervening once they are activated. That matters because a machine can make the decision to use lethal force, while critics question whether such attacks can reliably follow the rules of war.
The article gives the broad mechanism: armies are trying to use artificial intelligence to identify targets and strike them faster and more accurately. In general, AI systems can process information supplied to them and use it to classify or select targets; the article does not describe the sensors, software, or specific steps used by any weapon. It also says Stop Killer Robots reports such systems have been tested or deployed, including in Ukraine.
The central distinction is that a person does not intervene in the selection and attack once the system is activated. The article does not say that all weapons using AI are autonomous. It describes concerns about systems that can make and carry out those targeting decisions, and the talks sought rules for them.
How many countries are part of the talks, and what does it mean that a decision requires full consensus?
The article says 128 states party to the Convention on Conventional Weapons spent the week discussing whether to open negotiations on lethal autonomous weapons. They were also debating what terms and timeline any negotiations should have. The talks were held at the convention’s UN Review Conference in Geneva.
The key mechanism was full consensus: a decision required agreement from all participating states. In practical terms, a country that objects can prevent the group from adopting a decision, even if most countries support it. Russia’s objections and demands during the week were a focus of the reported deadlock, while one anonymous diplomat said only one delegation was actively opposing a negotiation-focused process.
By Friday afternoon, the states had not found a way past the impasse, and the conference was due to end that evening. One diplomat nevertheless said the camps had moved closer together on the need to negotiate something. The consensus rule therefore made agreement difficult, but the talks had not ended without signs of possible common ground.
What kind of agreement were countries debating, and how did they differ on whether it should be legally binding and how quickly to negotiate it?
Countries were debating whether to negotiate an agreement setting rules for lethal autonomous weapons, and what those rules and negotiations should look like. Most favoured a legally binding agreement. But countries with the technology, including the United States and Russia, were reluctant to have their hands tied. France supported a hybrid approach combining legally binding and non-binding measures.
They also disagreed about how long negotiations should take. The International Committee of the Red Cross and various non-governmental organisations wanted a one-year process. Washington proposed five years. Those differences meant that even countries supporting negotiations still had to settle the form and pace of the work.
The 128 states needed full consensus to make a decision. By Friday afternoon, they remained deadlocked, although an anonymous diplomat said countries were close to consensus on the need to negotiate something. The article gives no final agreement or timeline; it reports that the conference was due to end that evening and that a minority continued to block progress.
What risks can autonomous weapons pose for civilians, and why might it be difficult to hold someone accountable for their actions?
Critics say lethal autonomous weapons raise the risk of indiscriminate strikes and may not be used in line with the rules of war. That matters especially for civilians: the UN Secretary-General urged states to prevent the use and spread of what he called “merciless weapons” and to keep shielding civilians from the devastation of armed conflict.
The central accountability concern is that the system can select and attack a target without human intervention once activated. If an attack is unlawful or causes harm, critics say it may be unclear who is accountable for what the weapon did. The article identifies this as a concern, but does not explain how responsibility would be assigned in a particular case.
These risks formed part of the push for international rules. Most countries at the Geneva talks favoured a legally binding agreement, while some technology-holding states were reluctant to accept restrictions. The talks had not reached a decision by Friday afternoon. Without agreement, the article warns, it may become more difficult in the future to secure the rules campaigners say are needed.
What is the Convention on Conventional Weapons, and how does it provide a forum for countries to agree on rules for weapons?
The article calls the forum the United Nations’ Review Conference of the Convention on Conventional Weapons, or CCW. It brought together 128 states party to the convention to discuss whether to begin negotiations on rules for lethal autonomous weapons. The article does not explain the convention’s wider history or provisions.
As general background, the convention is formally known as the Convention on Certain Conventional Weapons. It is an international treaty framework through which participating countries can agree on restrictions or prohibitions concerning particular conventional weapons. Its structure allows states to discuss rules together and adopt them through agreement. This background is not detailed in the article.
In this case, the CCW provided the diplomatic setting for countries to debate whether to open a negotiation process, its terms, and its timeline. Full consensus was required for a decision, which made agreement difficult. The conference was due to end on Friday evening, but by mid-afternoon the states had not resolved the deadlock. The article reports no final set of rules.
How can AI use information from sensors to identify targets, and what is different about letting a machine make a lethal decision?
The article says armies are trying to use AI to identify and strike targets faster and more accurately. It does not specify which sensors provide information or how any particular system processes it. In general, sensors can gather information about the surroundings, and AI software can analyse that information to help identify objects or targets.
For example, a system might receive images from a camera and use software to classify what appears in them. That is a general illustration, not a detail reported about the weapons in this article. The article’s key point is that lethal autonomous weapons can select and attack targets once activated, without any human intervention.
Using AI to help identify a target is not the same as allowing a machine to decide to attack it. The latter removes human intervention from the selection and attack described in the article. Critics warn this can increase the risk of indiscriminate strikes, make accountability unclear, and create problems with following the rules of war. The talks sought rules for these systems.
This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.
Read more in the JupiteX app
Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.
Or read more news on the web