The United States and Russia forced the removal of key safeguards from a proposed United Nations framework governing lethal autonomous weapons — including a requirement that humans review targets selected by artificial intelligence — The Washington Post reported on Saturday.
The changes were made during negotiations held in Geneva earlier this month under the Convention on Certain Conventional Weapons. The talks marked the UN's most advanced attempt yet to set international rules for weapons that can identify, select and engage targets with limited or no human intervention.
According to the report, US and Russian diplomats spent nearly 15 hours revising the document on the final day of talks. The two delegations removed language requiring autonomous weapons to operate in a "predictable" and "reliable" manner, and deleted a provision calling for ethical considerations in how such systems are used. Another provision that would have required humans to review military targets identified by AI before an attack was also struck out.
The negotiations took place behind closed doors after UN cameras were switched off and civil society observers were asked to leave. Each of the US and Russian delegations fielded about 10 legal experts — nearly twice the number sent by many other countries, the Post reported. The US State Department, the Russian Foreign Ministry and the UN did not respond to the newspaper's requests for comment.
Human Rights Watch separately said the two countries had weakened provisions widely supported by other governments. The final text removed references to the design and development of autonomous weapons and narrowed the legal framework from international law in general to international humanitarian law. It also dropped references to explainability and traceability — safeguards meant to make it possible to understand how an AI system reached a targeting decision and to assign responsibility when something goes wrong.
The revised document still states that human control and judgement are required for compliance with international law, and it retains restrictions on systems that cannot operate within the rules of armed conflict. But Verity Coyle of Human Rights Watch warned that weaker rules could allow machines to make life-and-death decisions without effective human control — risking greater civilian harm, reduced accountability and the faster spread of automated warfare.
The Geneva process currently produces only non-binding recommendations, but the document could form the basis for negotiations on the world's first legally binding treaty on lethal autonomous weapons. Decisions under the convention are generally taken by consensus, giving major military powers the ability to block stronger provisions backed by a majority. Governments are expected to meet again in Geneva in November to decide whether three years of deliberations should move towards formal treaty negotiations.
The debate has gained urgency as AI embeds itself deeper into military planning. The Post previously reported that the US military used an Anthropic chatbot to help identify nearly 1,000 targets during the first 24 hours of its war with Iran. Separately, the Pentagon is revising its 2023 directive on autonomous weapons, which requires appropriate levels of human judgement over the use of force; President Trump ordered an update in June, but the revised version had not been released by late September.
Sources:
- US, Russia weaken UN killer-AI safeguards: Report — IANS
- Why Global Powers Removed Human Control From International Rules on Automated Weapons — Wall Street Next
The changes were made during negotiations held in Geneva earlier this month under the Convention on Certain Conventional Weapons. The talks marked the UN's most advanced attempt yet to set international rules for weapons that can identify, select and engage targets with limited or no human intervention.
According to the report, US and Russian diplomats spent nearly 15 hours revising the document on the final day of talks. The two delegations removed language requiring autonomous weapons to operate in a "predictable" and "reliable" manner, and deleted a provision calling for ethical considerations in how such systems are used. Another provision that would have required humans to review military targets identified by AI before an attack was also struck out.
The negotiations took place behind closed doors after UN cameras were switched off and civil society observers were asked to leave. Each of the US and Russian delegations fielded about 10 legal experts — nearly twice the number sent by many other countries, the Post reported. The US State Department, the Russian Foreign Ministry and the UN did not respond to the newspaper's requests for comment.
Human Rights Watch separately said the two countries had weakened provisions widely supported by other governments. The final text removed references to the design and development of autonomous weapons and narrowed the legal framework from international law in general to international humanitarian law. It also dropped references to explainability and traceability — safeguards meant to make it possible to understand how an AI system reached a targeting decision and to assign responsibility when something goes wrong.
The revised document still states that human control and judgement are required for compliance with international law, and it retains restrictions on systems that cannot operate within the rules of armed conflict. But Verity Coyle of Human Rights Watch warned that weaker rules could allow machines to make life-and-death decisions without effective human control — risking greater civilian harm, reduced accountability and the faster spread of automated warfare.
The Geneva process currently produces only non-binding recommendations, but the document could form the basis for negotiations on the world's first legally binding treaty on lethal autonomous weapons. Decisions under the convention are generally taken by consensus, giving major military powers the ability to block stronger provisions backed by a majority. Governments are expected to meet again in Geneva in November to decide whether three years of deliberations should move towards formal treaty negotiations.
The debate has gained urgency as AI embeds itself deeper into military planning. The Post previously reported that the US military used an Anthropic chatbot to help identify nearly 1,000 targets during the first 24 hours of its war with Iran. Separately, the Pentagon is revising its 2023 directive on autonomous weapons, which requires appropriate levels of human judgement over the use of force; President Trump ordered an update in June, but the revised version had not been released by late September.
Sources:
- US, Russia weaken UN killer-AI safeguards: Report — IANS
- Why Global Powers Removed Human Control From International Rules on Automated Weapons — Wall Street Next
