Skip to content
AI.info

The Pulse

U.S. and Russia Remove Human-Review Rule From U.N. AI Weapons Draft

U.S. and Russian diplomats removed a proposed human-review requirement and other safeguards from a U.N. draft on lethal autonomous weapons during September talks in Geneva. Delegates are due to meet again in November to consider whether neg

U.S. and Russia Remove Human-Review Rule From U.N. AI Weapons Draft

AI.info Team · Written with AI · fact-checked ·

Washington and Moscow pared back safeguards in Geneva

U.S. and Russian diplomats weakened a U.N. draft on lethal autonomous weapons during negotiations in Geneva in early September, removing proposed protections that would have required human review of AI-generated military targets before a strike. The changes also cut language calling for weapon systems to operate predictably and reliably, and a clause requiring ethical considerations when using AI weapons, The Washington Post reported on September 26.

The account draws on documents reviewed by the newspaper and three people familiar with the closed-door talks, who spoke anonymously. They said the final day of negotiations included roughly 15 hours of U.S. and Russian efforts to revise the text. One described the process as “death by a thousand paper cuts.”

The dispute is over who makes life-or-death decisions

For advocates of limits, the removed wording goes to the question of whether a person must remain involved when an autonomous system selects a target. Verity Coyle, deputy director for arms in Human Rights Watch’s crisis, conflict and arms division, warned that weak rules could leave machines making such decisions without human control.

“The practical impact is stark: more civilian harm, less accountability and a faster slide into riskier, automated warfare.”

Verity Coyle, Human Rights Watch

The United States has publicly rejected global controls on AI. Speaking at the United Nations on Tuesday, President Donald Trump said the country “totally rejects any attempt to construct a globalist scheme to control” artificial intelligence, which he said would be called “super intelligence.” The weapons negotiations, by contrast, remain nonbinding; a legally binding treaty would require agreement by member nations.

Small delegations faced a well-resourced push

The people familiar with the negotiations told The Post that the U.S. and Russian delegations each brought about 10 lawyers to the closed-door session, nearly twice the diplomatic presence of other delegations. U.N. cameras were off and civil society observers had been removed, they said, as the proposed edits moved quickly.

Nicole van Rooijen, executive director of the Stop Killer Robots coalition, said the negotiations require dozens of nations to reach consensus and that Washington and Moscow have substantial influence over the outcome. “They have the power to block any significant progress that is made by the vast majority of states who do want to move forward on a binding treaty,” she said.

The next test comes in November

Nations are scheduled to meet in Geneva in November to consider whether the talks can gain an official mandate or move toward a legally binding treaty. Richard Lennane, disarmament adviser for the legal team of the International Committee for the Red Cross, argued that treaties can shape how countries use weapons even when enforcement is not the only aim: “It’s not about punishment, it’s about setting norms and influencing behavior.”

The debate also overlaps with a U.S. review of the Pentagon’s own lethal autonomous weapons directive. A national security memorandum issued by Trump in June called for an updated directive within 90 days; as of late September, the government had not released a new version. The 2023 directive included requirements for human judgment in checking autonomous and semiautonomous weapons and for operators to follow the laws of war and applicable treaties.

Source

Related stories

Explore

More articles