US and Russia push for weaker oversight of AI-powered weapons at UN talks
US and Russian delegations at UN talks in Geneva secured the removal of human-oversight and ethics provisions from a draft global agreement on autonomous weapons. Rights groups warn this could let machines make life-and-death decisions without human control.
The Washington Post, citing diplomatic sources, reported on Saturday, September 26, that the United States and Russia jointly pushed for weakened controls over artificial intelligence-powered weapons. The report was also carried by Deutsche Welle, citing the Post.
The development took place at a UN meeting in early September in Geneva, where hundreds of diplomats discussed the first global agreement on the use of autonomous weapons capable of mass destruction. According to sources, on the final day of talks, US and Russian representatives spent 15 hours in a closed session finalizing the document's text.
Sources said the draft agreement's provisions on the reliability and predictability of such weapons were removed, along with requirements to account for ethical considerations when using AI systems for military purposes. Additionally, Russian and American diplomats succeeded in removing a provision requiring human verification of AI-identified military targets before a strike is carried out.
Closed-door process criticized
The closed session was held without UN cameras or civil society representatives present. Changes to the text were made so quickly that other participants could not keep pace with the editing process.
Verity Coyle, deputy director of the arms division at Human Rights Watch, said the weakened rules mean machines will be able to make life-and-death decisions without human oversight. She added this would lead to more civilian casualties, reduced accountability for war crimes, and a shift toward riskier, automated warfare.
Formally, the UN talks currently carry no binding legal force; countries are still determining whether the draft agreement could eventually become a formal mandate or international treaty. The next meeting is scheduled for November in Geneva.
In December 2024, the UN General Assembly adopted by an overwhelming majority its first resolution on lethal autonomous weapons, laying the groundwork for a UN-based global agreement. Only three countries voted against that resolution at the time — Russia, Belarus and North Korea.
The Washington Post noted that US forces have already used AI from the company Anthropic to detect and strike nearly 1,000 targets in the first 24 hours of the war with Iran. Israel has also made extensive use of AI in its military operations in the Gaza Strip.
